1 \input texinfo @c -*-texinfo-*-
3 @c Copyright (c) 1997-2014 Erez Zadok
4 @c Copyright (c) 1989 Jan-Simon Pendry
5 @c Copyright (c) 1989 Imperial College of Science, Technology & Medicine
6 @c Copyright (c) 1989 The Regents of the University of California.
7 @c All rights reserved.
9 @c This code is derived from software contributed to Berkeley by
10 @c Jan-Simon Pendry at Imperial College, London.
12 @c Redistribution and use in source and binary forms, with or without
13 @c modification, are permitted provided that the following conditions
15 @c 1. Redistributions of source code must retain the above copyright
16 @c notice, this list of conditions and the following disclaimer.
17 @c 2. Redistributions in binary form must reproduce the above copyright
18 @c notice, this list of conditions and the following disclaimer in the
19 @c documentation and/or other materials provided with the distribution.
20 @c 3. Neither the name of the University nor the names of its contributors
21 @c may be used to endorse or promote products derived from this software
22 @c without specific prior written permission.
24 @c THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 @c ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 @c IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 @c ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 @c FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 @c DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 @c OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 @c HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 @c LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 @c OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 @c File: am-utils/doc/am-utils.texi
38 @setfilename am-utils.info
42 @c info directory entry
43 @dircategory Administration
45 * Am-utils: (am-utils). The Amd automounter suite of utilities
48 @settitle Am-utils (4.4BSD Automounter Utilities)
49 @setchapternewpage odd
52 @title Am-utils (4.4BSD Automounter Utilities)
53 @subtitle For version @value{VERSION}, @value{UPDATED}
56 (Originally by Jan-Simon Pendry and Nick Williams)
59 Copyright @copyright{} 1997-2014 Erez Zadok
61 Copyright @copyright{} 1989 Jan-Simon Pendry
63 Copyright @copyright{} 1989 Imperial College of Science, Technology & Medicine
65 Copyright @copyright{} 1989 The Regents of the University of California.
69 Permission to copy this document, or any portion of it, as
70 necessary for use of this software is granted provided this
71 copyright notice and statement of permission are included.
75 @c Define a new index for options.
81 @c ################################################################
82 @node Top, License, , (DIR)
84 @b{Am-utils (4.4BSD Automounter Utilities) User Manual}
86 For version @value{VERSION}, @value{UPDATED}
90 (Originally by Jan-Simon Pendry and Nick Williams)
92 Copyright @copyright{} 1997-2014 Erez Zadok
94 Copyright @copyright{} 1989 Jan-Simon Pendry
96 Copyright @copyright{} 1989 Imperial College of Science, Technology & Medicine
98 Copyright @copyright{} 1989 The Regents of the University of California.
102 Permission to copy this document, or any portion of it, as
103 necessary for use of this software is granted provided this
104 copyright notice and statement of permission are included.
106 Am-utils is the 4.4BSD Automounter Tool Suite, which includes the Amd
107 automounter, the Amq query and control program, the Hlfsd daemon, and
108 other tools. This Info file describes how to use and understand the
109 tools within Am-utils.
113 * License:: Explains the terms and conditions for using
114 and distributing Am-utils.
115 * Distrib:: How to get the latest Am-utils distribution.
116 * AddInfo:: How to get additional information.
117 * Intro:: An introduction to Automounting concepts.
118 * History:: History of am-utils' development.
119 * Overview:: An overview of Amd.
120 * Supported Platforms:: Machines and Systems supported by Amd.
121 * Mount Maps:: Details of mount maps.
122 * Amd Command Line Options:: All the Amd command line options explained.
123 * Filesystem Types:: The different mount types supported by Amd.
124 * Amd Configuration File:: The amd.conf file syntax and meaning.
125 * Run-time Administration:: How to start, stop and control Amd.
126 * FSinfo:: The FSinfo filesystem management tool.
127 * Hlfsd:: The Home-Link Filesystem server.
128 * Assorted Tools:: Other tools which come with am-utils.
129 * Examples:: Some examples showing how Amd might be used.
130 * Internals:: Implementation details.
131 * Acknowledgments & Trademarks:: Legal Notes.
134 * Index:: An item for each concept.
140 This manual documents the use of the 4.4BSD automounter tool suite,
141 which includes @i{Amd}, @i{Amq}, @i{Hlfsd}, and other programs. This is
142 primarily a reference manual. While no tutorial exists, there are
143 examples available. @xref{Examples}.
145 This manual comes in two forms: the published form and the Info form.
146 The Info form is for on-line perusal with the INFO program which is
147 distributed along with GNU texinfo package (a version of which is
148 available for GNU Emacs).@footnote{GNU packages can be found in
149 @url{ftp://ftp.gnu.org/pub/gnu/}.} Both forms contain substantially
150 the same text and are generated from a common source file, which is
151 distributed with the @i{Am-utils} source.
154 @c ################################################################
155 @node License, Distrib, Top, Top
157 @cindex License Information
159 @i{Am-utils} is not in the public domain; it is copyrighted and there are
160 restrictions on its distribution.
162 Redistribution and use in source and binary forms, with or without
163 modification, are permitted provided that the following conditions are
169 Redistributions of source code must retain the above copyright notice,
170 this list of conditions and the following disclaimer.
173 Redistributions in binary form must reproduce the above copyright
174 notice, this list of conditions and the following disclaimer in the
175 documentation and/or other materials provided with the distribution.
178 Neither the name of the University nor the names of its contributors may
179 be used to endorse or promote products derived from this software
180 without specific prior written permission.
184 THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
185 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
186 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
187 PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS
188 BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
189 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
190 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
191 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
192 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
193 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
194 THE POSSIBILITY OF SUCH DAMAGE.
196 @c ################################################################
197 @node Distrib, AddInfo, License, Top
198 @unnumbered Source Distribution
199 @cindex Source code distribution
200 @cindex Obtaining the source code
202 The @i{Am-utils} home page is located in
204 @url{http://www.am-utils.org/}
207 You can get the latest distribution version of @i{Am-utils} from
209 @url{ftp://ftp.am-utils.org/pub/am-utils/am-utils.tar.gz}
212 Additional alpha, beta, and release distributions are available in
214 @url{ftp://ftp.am-utils.org/pub/am-utils/}.
217 Revision 5.2 was part of the 4.3BSD Reno distribution.
219 Revision 5.3bsdnet, a late alpha version of 5.3, was part
220 of the BSD network version 2 distribution
222 Revision 6.0 was made independently by
223 Erez Zadok at the Computer Science
224 Department of @uref{http://www.cs.columbia.edu/,Columbia University},
226 @uref{http://www.fsl.cs.sunysb.edu/docs/zadok-thesis-proposal/,PhD
227 thesis work}. Am-utils (especially version 6.1) continues to be
228 developed and maintained at the
229 @uref{http://www.cs.sunysb.edu/,Computer Science Department} of
230 @uref{http://www.stonybrook.edu/,Stony Brook University}, as a service
231 to the user community.
234 @xref{History}, for more details.
236 @c ################################################################
237 @node AddInfo, Intro, Distrib, Top
238 @unnumbered Getting Additional Information
239 @cindex Getting Additional Information
241 @unnumberedsec Bug Reports
244 Before reporting a bug, see if it is a known one in the
245 @uref{http://www.am-utils.org/docs/am-utils/BUGS.txt,bugs} file.
247 If you find a problem and hopefully you can reproduce it, please
248 describe it in detail and
249 @uref{https://bugzilla.filesystems.org/,submit a bug report} via
250 @uref{http://www.bugzilla.org/,Bugzilla}. Alternatively, you can send
251 your bug report to the ``am-utils'' list (see
252 @url{http://www.am-utils.org/} under ``Mailing Lists'') quoting the details
253 of the release and your configuration. These details can be obtained
254 by running the command @samp{amd -v}. It would greatly help if you
255 could provide a reproducible procedure for detecting the bug you are
258 Providing working patches is highly encouraged. Every patch
259 incorporated, however small, will get its author an honorable mention in
260 the @uref{http://www.am-utils.org/docs/am-utils/AUTHORS.txt,authors
263 @unnumberedsec Mailing Lists
264 @cindex Mailing lists
266 There are several mailing lists for people interested in keeping up-to-date
274 The users mailing list, @samp{am-utils} is for
278 announcements of alpha and beta releases of am-utils
280 reporting of bugs and patches
282 discussions of new features for am-utils
284 implementation and porting issues
287 To subscribe, visit @url{http://www.am-utils.org/} under ``Mailing
288 Lists.'' After subscribing, you can post a message to this list. To
289 avoid as much spam as possible, only subscribers to this list may post
292 Subscribers of @samp{am-utils} are most helpful if they have the time
293 and resources to test new and development versions of amd, on as many
294 different platforms as possible. They should also be prepared to
295 learn and use the GNU Autoconf, Automake, and Libtool packages, as
296 needed; and of course, become familiar with the complex code in the
297 am-utils package. In other words, subscribers on this list should
298 hopefully be able to contribute meaningfully to the development of
301 Note that this @samp{am-utils} list used to be called @samp{amd-dev}
302 before January 1st, 2004. Please use the new name, @samp{am-utils}.
305 The announcements mailing list, @samp{am-utils-announce} is for
306 announcements only (mostly new releases). To subscribe, visit
307 @url{http://www.am-utils.org/} under ``Mailing Lists.''
308 This list is read-only: only am-utils developers may post to it.
311 We distribute nightly CVS snapshots in
312 @url{ftp://ftp.am-utils.org/pub/am-utils/snapshots/daily/}. If you
313 like to get email notices of commits to the am-utils CVS repository,
314 subscribe to the CVS logs mailing list, @samp{am-utils-cvs} at
315 @url{http://www.am-utils.org/} under ``Mailing Lists.''
318 The older list which was used to user discussions, @samp{amd-workers},
319 is defunct as of January 2004. (Its last address was
320 @email{amd-workers AT majordomo.glue.umd.edu}.) Don't use
321 @samp{amd-workers}: use the newer, more active @samp{am-utils} list.
324 For completeness, there's a developers-only closed list called
325 @samp{am-utils-developers} (see @url{http://www.am-utils.org/} under
330 @unnumberedsec Am-utils Book
331 @cindex Am-utils book
333 @cindex Automounter book
336 @uref{http://www.cs.sunysb.edu/~ezk,Erez Zadok} wrote a
337 @uref{http://www.fsl.cs.sunysb.edu/docs/amd-book/,book}, titled @i{Linux NFS and
338 Automounter Administration}, ISBN 0-7821-2739-8, (Sybex, 2001). The
339 book is full of details and examples that go beyond what this manual
340 has. The book also covers NFS in great detail. Although the book is
341 geared toward Linux users, it is general enough for any Unix
342 administrator and contains specific sections for non-Linux systems.
344 @c ################################################################
345 @node Intro, History, AddInfo, Top
346 @unnumbered Introduction
349 An @dfn{automounter} maintains a cache of mounted filesystems.
350 Filesystems are mounted on demand when they are first referenced,
351 and unmounted after a period of inactivity.
353 @i{Amd} may be used as a replacement for Sun's automounter. The choice
354 of which filesystem to mount can be controlled dynamically with
355 @dfn{selectors}. Selectors allow decisions of the form ``hostname is
356 @var{this},'' or ``architecture is not @var{that}.'' Selectors may be
357 combined arbitrarily. @i{Amd} also supports a variety of filesystem
358 types, including NFS, UFS and the novel @dfn{program} filesystem. The
359 combination of selectors and multiple filesystem types allows identical
360 configuration files to be used on all machines thus reducing the
361 administrative overhead.
363 @i{Amd} ensures that it will not hang if a remote server goes down.
364 Moreover, @i{Amd} can determine when a remote server has become
365 inaccessible and then mount replacement filesystems as and when they
368 @i{Amd} contains no proprietary source code and has been ported to
369 numerous flavors of Unix.
371 @c ################################################################
372 @node History, Overview, Intro, Top
376 The @i{Amd} package has been without an official maintainer since 1992.
377 Several people have stepped in to maintain it unofficially. Most
378 notable were the `upl' (Unofficial Patch Level) releases of @i{Amd},
379 created by me (@uref{http://www.cs.sunysb.edu/~ezk,Erez Zadok}), and available from
380 @url{ftp://ftp.am-utils.org/pub/amd/}. The last such unofficial
381 release was `upl102'.
383 Through the process of patching and aging, it was becoming more and more
384 apparent that @i{Amd} was in much need of revitalizing. Maintaining
385 @i{Amd} had become a difficult task. I took it upon myself to cleanup
386 the code, so that it would be easier to port to new platforms, add new
387 features, keep up with the many new feature requests, and deal with the
388 never ending stream of bug reports.
390 I have been working on such a release of @i{Amd} on and off since
391 January of 1996. The new suite of tools is currently named "am-utils"
392 (AutoMounter Utilities), in line with GNU naming conventions, befitting
393 the contents of the package. In October of 1996 I had received enough
394 offers to help me with this task that I decided to make a mailing list
395 for this group of people. Around the same time, @i{Amd} had become a
396 necessary part of my PhD thesis work, resulting in more work performed
399 Am-utils version 6.0 was numbered with a major new release number to
400 distinguish it from the last official release of @i{Amd} (5.x). Many
401 new features have been added such as a GNU @code{configure} system, NFS
402 Version 3, a run-time configuration file (`amd.conf'), many new ports,
403 more scripts and programs, as well as numerous bug fixes. Another
404 reason for the new major release number was to alert users of am-utils
405 that user-visible interfaces may have changed. In order to make @i{Amd}
406 work well for the next 10 years, and be easier to maintain, it was
407 necessary to remove old or unused features, change various syntax files,
408 etc. However, great care was taken to ensure the maximum possible
409 backwards compatibility.
411 Am-utils version 6.1 has autofs support for Linux and Solaris 2.5+ as
412 @i{the} major new feature, in addition to several other minor new
413 features. The autofs support is completely transparent to the
414 end-user, aside from the fact that @code{/bin/pwd} now always returns
415 the correct amd-ified path. The administrator can easily switch
416 between NFS and autofs mounts by changing one parameter in
417 @code{amd.conf}. Autofs support and maintenance was developed in
418 conjunction with @email{ionut AT badula.org,Ion Badulescu}.
420 @c ################################################################
421 @node Overview, Supported Platforms, History, Top
424 @i{Amd} maintains a cache of mounted filesystems. Filesystems are
425 @dfn{demand-mounted} when they are first referenced, and unmounted after
426 a period of inactivity. @i{Amd} may be used as a replacement for Sun's
427 @b{automount}(8) program. It contains no proprietary source code and
428 has been ported to numerous flavors of Unix. @xref{Supported
431 @i{Amd} was designed as the basis for experimenting with filesystem
432 layout and management. Although @i{Amd} has many direct applications it
433 is loaded with additional features which have little practical use. At
434 some point the infrequently used components may be removed to streamline
435 the production system.
437 @i{Amd} supports the notion of @dfn{replicated} filesystems by evaluating
438 each member of a list of possible filesystem locations one by one.
439 @i{Amd} checks that each cached mapping remains valid. Should a mapping be
440 lost -- such as happens when a fileserver goes down -- @i{Amd} automatically
441 selects a replacement should one be available.
445 * Filesystems and Volumes::
448 * Operational Principles::
449 * Mounting a Volume::
450 * Automatic Unmounting::
452 * Non-blocking Operation::
455 @node Fundamentals, Filesystems and Volumes, Overview, Overview
456 @comment node-name, next, previous, up
457 @section Fundamentals
458 @cindex Automounter fundamentals
460 The fundamental concept behind @i{Amd} is the ability to separate the
461 name used to refer to a file from the name used to refer to its physical
462 storage location. This allows the same files to be accessed with the
463 same name regardless of where in the network the name is used. This is
464 very different from placing @file{/n/hostname} in front of the pathname
465 since that includes location dependent information which may change if
466 files are moved to another machine.
468 By placing the required mappings in a centrally administered database,
469 filesystems can be re-organized without requiring changes to
470 configuration files, shell scripts and so on.
472 @node Filesystems and Volumes, Volume Naming, Fundamentals, Overview
473 @comment node-name, next, previous, up
474 @section Filesystems and Volumes
480 @i{Amd} views the world as a set of fileservers, each containing one or
481 more filesystems where each filesystem contains one or more
482 @dfn{volumes}. Here the term @dfn{volume} is used to refer to a
483 coherent set of files such as a user's home directory or a @TeX{}
486 In order to access the contents of a volume, @i{Amd} must be told in
487 which filesystem the volume resides and which host owns the filesystem.
488 By default the host is assumed to be local and the volume is assumed to
489 be the entire filesystem. If a filesystem contains more than one
490 volume, then a @dfn{sublink} is used to refer to the sub-directory
491 within the filesystem where the volume can be found.
493 @node Volume Naming, Volume Binding, Filesystems and Volumes, Overview
494 @comment node-name, next, previous, up
495 @section Volume Naming
497 @cindex Network-wide naming
498 @cindex Replicated volumes
499 @cindex Duplicated volumes
500 @cindex Replacement volumes
502 Volume names are defined to be unique across the entire network. A
503 volume name is the pathname to the volume's root as known by the users
504 of that volume. Since this name uniquely identifies the volume
505 contents, all volumes can be named and accessed from each host, subject
506 to administrative controls.
508 Volumes may be replicated or duplicated. Replicated volumes contain
509 identical copies of the same data and reside at two or more locations in
510 the network. Each of the replicated volumes can be used
511 interchangeably. Duplicated volumes each have the same name but contain
512 different, though functionally identical, data. For example,
513 @samp{/vol/tex} might be the name of a @TeX{} distribution which varied
514 for each machine architecture.@refill
516 @i{Amd} provides facilities to take advantage of both replicated and
517 duplicated volumes. Configuration options allow a single set of
518 configuration data to be shared across an entire network by taking
519 advantage of replicated and duplicated volumes.
521 @i{Amd} can take advantage of replacement volumes by mounting them as
522 required should an active fileserver become unavailable.
524 @node Volume Binding, Operational Principles, Volume Naming, Overview
525 @comment node-name, next, previous, up
526 @section Volume Binding
527 @cindex Volume binding
528 @cindex Unix namespace
530 @cindex Binding names to filesystems
532 Unix implements a namespace of hierarchically mounted filesystems. Two
533 forms of binding between names and files are provided. A @dfn{hard
534 link} completes the binding when the name is added to the filesystem. A
535 @dfn{soft link} delays the binding until the name is accessed. An
536 @dfn{automounter} adds a further form in which the binding of name to
537 filesystem is delayed until the name is accessed.@refill
539 The target volume, in its general form, is a tuple (host, filesystem,
540 sublink) which can be used to name the physical location of any volume
543 When a target is referenced, @i{Amd} ignores the sublink element and
544 determines whether the required filesystem is already mounted. This is
545 done by computing the local mount point for the filesystem and checking
546 for an existing filesystem mounted at the same place. If such a
547 filesystem already exists then it is assumed to be functionally
548 identical to the target filesystem. By default there is a one-to-one
549 mapping between the pair (host, filesystem) and the local mount point so
550 this assumption is valid.
552 @node Operational Principles, Mounting a Volume, Volume Binding, Overview
553 @comment node-name, next, previous, up
554 @section Operational Principles
555 @cindex Operational principles
557 @i{Amd} operates by introducing new mount points into the namespace.
558 These are called @dfn{automount} points. The kernel sees these
559 automount points as NFS filesystems being served by @i{Amd}. Having
560 attached itself to the namespace, @i{Amd} is now able to control the
561 view the rest of the system has of those mount points. RPC calls are
562 received from the kernel one at a time.
564 When a @dfn{lookup} call is received @i{Amd} checks whether the name is
565 already known. If it is not, the required volume is mounted. A
566 symbolic link pointing to the volume root is then returned. Once the
567 symbolic link is returned, the kernel will send all other requests
568 direct to the mounted filesystem.
570 If a volume is not yet mounted, @i{Amd} consults a configuration
571 @dfn{mount-map} corresponding to the automount point. @i{Amd} then
572 makes a runtime decision on what and where to mount a filesystem based
573 on the information obtained from the map.
575 @i{Amd} does not implement all the NFS requests; only those relevant
576 to name binding such as @dfn{lookup}, @dfn{readlink} and @dfn{readdir}.
577 Some other calls are also implemented but most simply return an error
578 code; for example @dfn{mkdir} always returns ``read-only filesystem''.
580 @node Mounting a Volume, Automatic Unmounting, Operational Principles, Overview
581 @comment node-name, next, previous, up
582 @section Mounting a Volume
583 @cindex Mounting a volume
584 @cindex Location lists
585 @cindex Alternate locations
586 @cindex Mount retries
587 @cindex Background mounts
589 Each automount point has a corresponding mount map. The mount map
590 contains a list of key--value pairs. The key is the name of the volume
591 to be mounted. The value is a list of locations describing where the
592 filesystem is stored in the network. In the source for the map the
593 value would look like
596 location1 location2 @dots{} locationN
599 @i{Amd} examines each location in turn. Each location may contain
600 @dfn{selectors} which control whether @i{Amd} can use that location.
601 For example, the location may be restricted to use by certain hosts.
602 Those locations which cannot be used are ignored.
604 @i{Amd} attempts to mount the filesystem described by each remaining
605 location until a mount succeeds or @i{Amd} can no longer proceed. The
606 latter can occur in three ways:
610 If none of the locations could be used, or if all of the locations
611 caused an error, then the last error is returned.
614 If a location could be used but was being mounted in the background then
615 @i{Amd} marks that mount as being ``in progress'' and continues with
616 the next request; no reply is sent to the kernel.
619 Lastly, one or more of the mounts may have been @dfn{deferred}. A mount
620 is deferred if extra information is required before the mount can
621 proceed. When the information becomes available the mount will take
622 place, but in the mean time no reply is sent to the kernel. If the
623 mount is deferred, @i{Amd} continues to try any remaining locations.
626 Once a volume has been mounted, @i{Amd} establishes a @dfn{volume
627 mapping} which is used to satisfy subsequent requests.@refill
629 @node Automatic Unmounting, Keep-alives, Mounting a Volume, Overview
630 @comment node-name, next, previous, up
631 @section Automatic Unmounting
633 To avoid an ever increasing number of filesystem mounts, @i{Amd} removes
634 volume mappings which have not been used recently. A time-to-live
635 interval is associated with each mapping and when that expires the
636 mapping is removed. When the last reference to a filesystem is removed,
637 that filesystem is unmounted. If the unmount fails, for example the
638 filesystem is still busy, the mapping is re-instated and its
639 time-to-live interval is extended. The global default for this grace
640 period is controlled by the @code{-w} command-line option (@pxref{-w
641 Option, -w}) or the @i{amd.conf} parameter @samp{dismount_interval}
642 (@pxref{dismount_interval Parameter}). It is also possible to set this
643 value on a per-mount basis (@pxref{opts Option, opts, opts}).
645 Filesystems can be forcefully timed out using the @i{Amq} command.
646 @xref{Run-time Administration}. Note that on new enough systems that
647 support forced unmounts, such as Linux, @i{Amd} can try to use the
648 @b{umount2}(2) system call to force the unmount, if the regular
649 @b{umount}(2) system call failed in a way that indicates that the
650 mount point is hung or stale. @xref{forced_unmounts Parameter}.
652 @node Keep-alives, Non-blocking Operation, Automatic Unmounting, Overview
653 @comment node-name, next, previous, up
656 @cindex Server crashes
659 Use of some filesystem types requires the presence of a server on
660 another machine. If a machine crashes then it is of no concern to
661 processes on that machine that the filesystem is unavailable. However,
662 to processes on a remote host using that machine as a fileserver this
663 event is important. This situation is most widely recognized when an
664 NFS server crashes and the behavior observed on client machines is that
665 more and more processes hang. In order to provide the possibility of
666 recovery, @i{Amd} implements a @dfn{keep-alive} interval timer for some
667 filesystem types. Currently only NFS makes use of this service.
669 The basis of the NFS keep-alive implementation is the observation that
670 most sites maintain replicated copies of common system data such as
671 manual pages, most or all programs, system source code and so on. If
672 one of those servers goes down it would be reasonable to mount one of
673 the others as a replacement.
675 The first part of the process is to keep track of which fileservers are
676 up and which are down. @i{Amd} does this by sending RPC requests to the
677 servers' NFS @code{NullProc} and checking whether a reply is returned.
678 While the server state is uncertain the requests are re-transmitted at
679 three second intervals and if no reply is received after four attempts
680 the server is marked down. If a reply is received the fileserver is
681 marked up and stays in that state for 30 seconds at which time another
682 NFS ping is sent. This interval is configurable and can even be
683 turned off using the @i{ping} option. @xref{opts Option}.
685 Once a fileserver is marked down, requests continue to be sent every 30
686 seconds in order to determine when the fileserver comes back up. During
687 this time any reference through @i{Amd} to the filesystems on that
688 server fail with the error ``Operation would block''. If a replacement
689 volume is available then it will be mounted, otherwise the error is
690 returned to the user.
692 @c @i{Amd} keeps track of which servers are up and which are down.
693 @c It does this by sending RPC requests to the servers' NFS {\sc NullProc} and
694 @c checking whether a reply is returned. If no replies are received after a
695 @c short period, @i{Amd} marks the fileserver @dfn{down}.
696 @c RPC requests continue to be sent so that it will notice when a fileserver
698 @c ICMP echo packets \cite{rfc:icmp} are not used because it is the availability
699 @c of the NFS service that is important, not the existence of a base kernel.
700 @c Whenever a reference to a fileserver which is down is made via @i{Amd}, an alternate
701 @c filesystem is mounted if one is available.
703 Although this action does not protect user files, which are unique on
704 the network, or processes which do not access files via @i{Amd} or
705 already have open files on the hung filesystem, it can prevent most new
706 processes from hanging.
708 @c With a suitable combination of filesystem management and mount-maps,
709 @c machines can be protected against most server downtime. This can be
710 @c enhanced by allocating boot-servers dynamically which allows a diskless
711 @c workstation to be quickly restarted if necessary. Once the root filesystem
712 @c is mounted, @i{Amd} can be started and allowed to mount the remainder of
713 @c the filesystem from whichever fileservers are available.
715 @node Non-blocking Operation, , Keep-alives, Overview
716 @comment node-name, next, previous, up
717 @section Non-blocking Operation
718 @cindex Non-blocking operation
719 @cindex Multiple-threaded server
722 Since there is only one instance of @i{Amd} for each automount point,
723 and usually only one instance on each machine, it is important that it
724 is always available to service kernel calls. @i{Amd} goes to great
725 lengths to ensure that it does not block in a system call. As a last
726 resort @i{Amd} will fork before it attempts a system call that may block
727 indefinitely, such as mounting an NFS filesystem. Other tasks such as
728 obtaining filehandle information for an NFS filesystem, are done using a
729 purpose built non-blocking RPC library which is integrated with
730 @i{Amd}'s task scheduler. This library is also used to implement NFS
731 keep-alives (@pxref{Keep-alives}).
733 Whenever a mount is deferred or backgrounded, @i{Amd} must wait for it
734 to complete before replying to the kernel. However, this would cause
735 @i{Amd} to block waiting for a reply to be constructed. Rather than do
736 this, @i{Amd} simply @dfn{drops} the call under the assumption that the
737 kernel RPC mechanism will automatically retry the request.
739 @c ################################################################
740 @node Supported Platforms, Mount Maps, Overview, Top
741 @comment node-name, next, previous, up
742 @chapter Supported Platforms
743 @cindex Supported Platforms
744 @cindex shared libraries
745 @cindex NFS V.3 support
747 @i{Am-utils} has been ported to a wide variety of machines and operating
748 systems. @i{Am-utils}'s code works for little-endian and big-endian
749 machines, as well as 32 bit and 64 bit architectures. Furthermore, when
750 @i{Am-utils} ports to an Operating System on one architecture, it is generally
751 readily portable to the same Operating System on all platforms on which
754 See the @file{INSTALL} in the distribution for more specific details on
755 building and/or configuring for some systems.
757 @c ################################################################
758 @node Mount Maps, Amd Command Line Options, Supported Platforms, Top
759 @comment node-name, next, previous, up
762 @cindex Automounter configuration maps
763 @cindex Mount information
765 @i{Amd} has no built-in knowledge of machines or filesystems.
766 External @dfn{mount-maps} are used to provide the required information.
767 Specifically, @i{Amd} needs to know when and under what conditions it
768 should mount filesystems.
770 The map entry corresponding to the requested name contains a list of
771 possible locations from which to resolve the request. Each location
772 specifies filesystem type, information required by that filesystem (for
773 example the block special device in the case of UFS), and some
774 information describing where to mount the filesystem (@pxref{fs Option}). A
775 location may also contain @dfn{selectors} (@pxref{Selectors}).@refill
783 @node Map Types, Key Lookup, Mount Maps, Mount Maps
784 @comment node-name, next, previous, up
786 @cindex Mount map types
788 @cindex Configuration map types
789 @cindex Types of mount map
790 @cindex Types of configuration map
791 @cindex Determining the map type
793 A mount-map provides the run-time configuration information to @i{Amd}.
794 Maps can be implemented in many ways. Some of the forms supported by
795 @i{Amd} are regular files, ndbm databases, NIS maps, the @dfn{Hesiod}
796 name server, and even the password file.
798 A mount-map @dfn{name} is a sequence of characters. When an automount
799 point is created a handle on the mount-map is obtained. For each map
800 type configured, @i{Amd} attempts to reference the map of the
801 appropriate type. If a map is found, @i{Amd} notes the type for future
802 use and deletes the reference, for example closing any open file
803 descriptors. The available maps are configured when @i{Amd} is built
804 and can be displayed by running the command @samp{amd -v}.
806 When using an @i{Amd} configuration file (@pxref{Amd Configuration File})
807 and the keyword @samp{map_type} (@pxref{map_type Parameter}), you may
808 force the map used to any type.
810 By default, @i{Amd} caches data in a mode dependent on the type of map.
811 This is the same as specifying @samp{cache:=mapdefault} and selects a
812 suitable default cache mode depending on the map type. The individual
813 defaults are described below. The @var{cache} option can be specified
814 on automount points to alter the caching behavior (@pxref{Automount
817 The following map types have been implemented, though some are not
818 available on all machines. Run the command @samp{amd -v} to obtain a
819 list of map types configured on your machine.
833 @c ----------------------------------------------------------------
834 @node File maps, ndbm maps, Map Types, Map Types
835 @comment node-name, next, previous, up
836 @subsection File maps
838 @cindex Flat file maps
839 @cindex File map syntactic conventions
841 When @i{Amd} searches a file for a map entry it does a simple scan of
842 the file and supports both comments and continuation lines.
844 Continuation lines are indicated by a backslash character (@samp{\}) as
845 the last character of a line in the file. The backslash, newline character
846 @emph{and any leading white space on the following line} are discarded. A maximum
847 line length of 2047 characters is enforced after continuation lines are read
848 but before comments are stripped. Each line must end with
849 a newline character; that is newlines are terminators, not separators.
850 The following examples illustrate this:
857 specifies @emph{three} locations, and is identical to
870 specifies only @emph{two} locations, and is identical to
876 After a complete line has been read from the file, including
877 continuations, @i{Amd} determines whether there is a comment on the
878 line. A comment begins with a hash (``@samp{#}'') character and
879 continues to the end of the line. There is no way to escape or change
880 the comment lead-in character.
882 Note that continuation lines and comment support @dfn{only} apply to
883 file maps, or ndbm maps built with the @code{mk-amd-map} program.
885 When caching is enabled, file maps have a default cache mode of
886 @code{all} (@pxref{Automount Filesystem}).
888 @c ----------------------------------------------------------------
889 @node ndbm maps, NIS maps, File maps, Map Types
890 @comment node-name, next, previous, up
891 @subsection ndbm maps
894 An ndbm map may be used as a fast access form of a file map. The program,
895 @code{mk-amd-map}, converts a normal map file into an ndbm database.
896 This program supports the same continuation and comment conventions that
897 are provided for file maps. Note that ndbm format files may @emph{not}
898 be sharable across machine architectures. The notion of speed generally
899 only applies to large maps; a small map, less than a single disk block,
900 is almost certainly better implemented as a file map.
902 ndbm maps have a default cache mode of @samp{all}
903 (@pxref{Automount Filesystem}).
905 @c ----------------------------------------------------------------
906 @node NIS maps, NIS+ maps, ndbm maps, Map Types
907 @comment node-name, next, previous, up
909 @cindex NIS (YP) maps
911 When using NIS (formerly YP), an @i{Amd} map is implemented directly
912 by the underlying NIS map. Comments and continuation lines are
913 @emph{not} supported in the automounter and must be stripped when
914 constructing the NIS server's database.
916 NIS maps have a default cache mode of @code{all} (@pxref{Automount
919 The following rule illustrates what could be added to your NIS @file{Makefile},
920 in this case causing the @file{amd.home} map to be rebuilt:
922 $(YPTSDIR)/amd.home.time: $(ETCDIR)/amd.home
923 -@@sed -e "s/#.*$$//" -e "/^$$/d" $(ETCDIR)/amd.home | \
925 for (i = 1; i <= NF; i++) \
927 if (substr($$i, length($$i), 1) == "\\") \
928 printf("%s", substr($$i, 1, length($$i) - 1)); \
930 printf("%s\n", $$i); \
933 printf("%s ", $$i); \
935 $(MAKEDBM) - $(YPDBDIR)/amd.home; \
936 touch $(YPTSDIR)/amd.home.time; \
937 echo "updated amd.home"; \
938 if [ ! $(NOPUSH) ]; then \
939 $(YPPUSH) amd.home; \
940 echo "pushed amd.home"; \
946 Here @code{$(YPTSDIR)} contains the time stamp files, and
947 @code{$(YPDBDIR)} contains the dbm format NIS files.
949 @c ----------------------------------------------------------------
950 @node NIS+ maps, Hesiod maps, NIS maps, Map Types
951 @comment node-name, next, previous, up
952 @subsection NIS+ maps
955 NIS+ maps do not support cache mode @samp{all} and, when caching is
956 enabled, have a default cache mode of @samp{inc}.
958 XXX: FILL IN WITH AN EXAMPLE.
960 @c ----------------------------------------------------------------
961 @node Hesiod maps, Password maps, NIS+ maps, Map Types
962 @comment node-name, next, previous, up
963 @subsection Hesiod maps
966 When the map name begins with the string @samp{hesiod.} lookups are made
967 using the @dfn{Hesiod} name server. The string following the dot is
968 used as a name qualifier and is prepended with the key being located.
969 The entire string is then resolved in the @code{automount} context, or
970 the @i{amd.conf} parameter @samp{hesiod_base} (@pxref{hesiod_base
971 Parameter}). For example, if the key is @samp{jsp} and map name is
972 @samp{hesiod.homes} then @dfn{Hesiod} is asked to resolve
973 @samp{jsp.homes.automount}.
975 Hesiod maps do not support cache mode @samp{all} and, when caching is
976 enabled, have a default cache mode of @samp{inc} (@pxref{Automount
979 The following is an example of a @dfn{Hesiod} map entry:
982 jsp.homes.automount HS TXT "rfs:=/home/charm;rhost:=charm;sublink:=jsp"
983 njw.homes.automount HS TXT "rfs:=/home/dylan/dk2;rhost:=dylan;sublink:=njw"
986 @c ----------------------------------------------------------------
987 @node Password maps, Union maps, Hesiod maps, Map Types
988 @comment node-name, next, previous, up
989 @subsection Password maps
990 @cindex Password file maps
991 @cindex /etc/passwd maps
992 @cindex User maps, automatic generation
993 @cindex Automatic generation of user maps
994 @cindex Using the password file as a map
996 The password map support is unlike the four previous map types. When
997 the map name is the string @file{/etc/passwd} @i{Amd} can lookup a user
998 name in the password file and re-arrange the home directory field to
999 produce a usable map entry.
1001 @i{Amd} assumes the home directory has the format
1002 `@t{/}@i{anydir}@t{/}@i{dom1}@t{/../}@i{domN}@t{/}@i{login}'.
1003 @c @footnote{This interpretation is not necessarily exactly what you want.}
1004 It breaks this string into a map entry where @code{$@{rfs@}} has the
1005 value `@t{/}@i{anydir}@t{/}@i{domN}', @code{$@{rhost@}} has the value
1006 `@i{domN}@t{.}@i{...}@t{.}@i{dom1}', and @code{$@{sublink@}} has the
1007 value @i{login}.@refill
1009 Thus if the password file entry was
1015 the map entry used by @i{Amd} would be
1018 rfs:=/home/achilles;rhost:=achilles;sublink:=jsp
1021 Similarly, if the password file entry was
1027 the map entry used by @i{Amd} would be
1030 rfs:=/home/sugar;rhost:=sugar.cc;sublink:=mhj
1033 @c ----------------------------------------------------------------
1034 @node Union maps, LDAP maps, Password maps, Map Types
1035 @comment node-name, next, previous, up
1036 @subsection Union maps
1037 @cindex Union file maps
1039 The union map support is provided specifically for use with the union
1040 filesystem, @pxref{Union Filesystem}.
1042 It is identified by the string @samp{union:} which is followed by a
1043 colon separated list of directories. The directories are read in order,
1044 and the names of all entries are recorded in the map cache. Later
1045 directories take precedence over earlier ones. The union filesystem
1046 type then uses the map cache to determine the union of the names in all
1049 @c ----------------------------------------------------------------
1050 @node LDAP maps, Executable maps, Union maps, Map Types
1051 @comment node-name, next, previous, up
1052 @subsection LDAP maps
1054 @cindex Lightweight Directory Access Protocol
1056 LDAP (Lightweight Directory Access Protocol) maps do not support cache
1057 mode @samp{all} and, when caching is enabled, have a default cache mode
1060 For example, an @i{Amd} map @samp{amd.home} that looks as follows:
1063 /defaults opts:=rw,intr;type:=link
1065 zing -rhost:=shekel \
1067 host!=shekel;type:=nfs
1070 when converted to LDAP (@pxref{amd2ldif}), will result in the following
1073 $ amd2ldif amd.home CUCS < amd.home
1074 dn: cn=amdmap timestamp, CUCS
1075 cn : amdmap timestamp
1076 objectClass : amdmapTimestamp
1077 amdmapTimestamp: 873071363
1079 dn: cn=amdmap amd.home[/defaults], CUCS
1080 cn : amdmap amd.home[/defaults]
1081 objectClass : amdmap
1082 amdmapName : amd.home
1083 amdmapKey : /defaults
1084 amdmapValue : opts:=rw,intr;type:=link
1086 dn: cn=amdmap amd.home[], CUCS
1087 cn : amdmap amd.home[]
1088 objectClass : amdmap
1089 amdmapName : amd.home
1093 dn: cn=amdmap amd.home[zing], CUCS
1094 cn : amdmap amd.home[zing]
1095 objectClass : amdmap
1096 amdmapName : amd.home
1098 amdmapValue : -rhost:=shekel host==shekel host!=shekel;type:=nfs
1101 @c ----------------------------------------------------------------
1102 @node Executable maps, , LDAP maps, Map Types
1103 @comment node-name, next, previous, up
1104 @subsection Executable maps
1105 @cindex Executable maps
1107 An executable map is a dynamic map in which the keys and values for
1108 the maps are generated on the fly by a program or script. The program
1109 is expected to take a single parameter argument which is the key to
1110 lookup. If the key is found, the program should print on stdout the
1111 key-value pair that were found; if the key was not found, nothing
1112 should be printed out. Below is an sample of such a map script:
1116 # executable map example
1119 echo "/defaults type:=nfs;rfs:=filer"
1122 echo "a type:=nfs;fs:=/tmp"
1125 echo "b type:=link;fs:=/usr/local"
1127 * ) # no match, echo nothing
1132 @xref{exec_map_timeout Parameter}.
1134 @c ----------------------------------------------------------------
1136 @c ----------------------------------------------------------------
1137 @node Key Lookup, Location Format, Map Types, Mount Maps
1138 @comment node-name, next, previous, up
1139 @section How keys are looked up
1142 @cindex Looking up keys
1143 @cindex How keys are looked up
1144 @cindex Wildcards in maps
1146 The key is located in the map whose type was determined when the
1147 automount point was first created. In general the key is a pathname
1148 component. In some circumstances this may be modified by variable
1149 expansion (@pxref{Variable Expansion}) and prefixing. If the automount
1150 point has a prefix, specified by the @var{pref} option, then that is
1151 prepended to the search key before the map is searched.
1153 If the map cache is a @samp{regexp} cache then the key is treated as an
1154 egrep-style regular expression, otherwise a normal string comparison is
1157 If the key cannot be found then a @dfn{wildcard} match is attempted.
1158 @i{Amd} repeatedly strips the basename from the key, appends @samp{/*} and
1159 attempts a lookup. Finally, @i{Amd} attempts to locate the special key @samp{*}.
1161 For example, the following sequence would be checked if @file{home/dylan/dk2} was
1171 At any point when a wildcard is found, @i{Amd} proceeds as if an exact
1172 match had been found and the value field is then used to resolve the
1173 mount request, otherwise an error code is propagated back to the kernel.
1174 (@pxref{Filesystem Types}).@refill
1176 @node Location Format, , Key Lookup, Mount Maps
1177 @comment node-name, next, previous, up
1178 @section Location Format
1179 @cindex Location format
1180 @cindex Map entry format
1181 @cindex How locations are parsed
1183 The value field from the lookup provides the information required to
1184 mount a filesystem. The information is parsed according to the syntax
1189 @i{location-selection}
1190 @i{location-list} @i{white-space} @t{||} @i{white-space} @i{location-selection}
1191 @i{location-selection}:
1193 @i{location-selection} @i{white-space} @i{location}
1196 @t{-}@i{location-info}
1200 @i{location-info}@t{;}@i{sel-or-opt}
1206 selector@t{==}@i{value}
1207 selector@t{!=}@i{value}
1209 option@t{:=}@i{value}
1215 Note that unquoted whitespace is not allowed in a location description.
1216 White space is only allowed, and is mandatory, where shown with non-terminal
1219 A @dfn{location-selection} is a list of possible volumes with which to
1220 satisfy the request. Each @dfn{location-selection} is tried
1221 sequentially, until either one succeeds or all fail. This, by the
1222 way, is different from the historically documented behavior, which
1223 claimed (falsely, at least for last 3 years) that @i{Amd} would
1224 attempt to mount all @dfn{location-selection}s in parallel and the
1225 first one to succeed would be used.
1227 @dfn{location-selection}s are optionally separated by the @samp{||}
1228 operator. The effect of this operator is to prevent use of
1229 location-selections to its right if any of the location-selections on
1230 its left were selected, whether or not any of them were successfully
1231 mounted (@pxref{Selectors}).@refill
1233 The location-selection, and singleton @dfn{location-list},
1234 @samp{type:=ufs;dev:=/dev/xd1g} would inform @i{Amd} to mount a UFS
1235 filesystem from the block special device @file{/dev/xd1g}.
1237 The @dfn{sel-or-opt} component is either the name of an option required
1238 by a specific filesystem, or it is the name of a built-in, predefined
1239 selector such as the architecture type. The value may be quoted with
1240 double quotes @samp{"}, for example
1241 @samp{type:="ufs";dev:="/dev/xd1g"}. These quotes are stripped when the
1242 value is parsed and there is no way to get a double quote into a value
1243 field. Double quotes are used to get white space into a value field,
1244 which is needed for the program filesystem (@pxref{Program Filesystem}).@refill
1248 * Variable Expansion::
1253 @node Map Defaults, Variable Expansion, Location Format, Location Format
1254 @comment node-name, next, previous, up
1255 @subsection Map Defaults
1256 @cindex Map defaults
1257 @cindex How to set default map parameters
1258 @cindex Setting default map parameters
1260 A location beginning with a dash @samp{-} is used to specify default
1261 values for subsequent locations. Any previously specified defaults in
1262 the location-list are discarded. The default string can be empty in
1263 which case no defaults apply.
1265 The location @samp{-fs:=/mnt;opts:=ro} would set the local mount point
1266 to @file{/mnt} and cause mounts to be read-only by default. Defaults
1267 specified this way are appended to, and so override, any global map
1268 defaults given with @samp{/defaults}).
1271 @c A @samp{/defaults} value @dfn{gdef} and a location list
1273 @c $@samp{-}@dfn{def}_a $\verb*+ +$ @dfn{loc}_{a_1} $\verb*+ +$ @dfn{loc}_{a_2} $\verb*+ +$ @samp{-}@dfn{def}_b $\verb*+ +$ @dfn{loc}_{b_1} \ldots$
1277 @c $@samp{-}@dfn{gdef}@samp{;}@dfn{def}_a $\verb*+ +$ @dfn{loc}_{a_1} $\verb*+ +$ @dfn{loc}_{a_2} $\verb*+ +$ @samp{-}@dfn{gdef}@samp{;}@dfn{def}_b $\verb*+ +$ @dfn{loc}_{b_1} \ldots$
1279 @c which is equivalent to
1281 @c $@dfn{gdef}@samp{;}@dfn{def}_a@samp{;}@dfn{loc}_{a_1} $\verb*+ +$@dfn{gdef}@samp{;}@dfn{def}_a@samp{;}@dfn{loc}_{a_2} $\verb*+ +$@dfn{gdef}@samp{;}@dfn{def}_b@samp{;}@dfn{loc}_{b_1} \ldots$
1284 @node Variable Expansion, Selectors, Map Defaults, Location Format
1285 @comment node-name, next, previous, up
1286 @subsection Variable Expansion
1287 @cindex Variable expansion
1288 @cindex How variables are expanded
1289 @cindex Pathname operators
1290 @cindex Domain stripping
1291 @cindex Domainname operators
1292 @cindex Stripping the local domain name
1293 @cindex Environment variables
1294 @cindex How to access environment variables in maps
1296 To allow generic location specifications @i{Amd} does variable expansion
1297 on each location and also on some of the option strings. Any option or
1298 selector appearing in the form @code{$@dfn{var}} is replaced by the
1299 current value of that option or selector. For example, if the value of
1300 @code{$@{key@}} was @samp{bin}, @code{$@{autodir@}} was @samp{/a} and
1301 @code{$@{fs@}} was `@t{$@{autodir@}}@t{/local/}@t{$@{key@}}' then
1302 after expansion @code{$@{fs@}} would have the value @samp{/a/local/bin}.
1303 Any environment variable can be accessed in a similar way.@refill
1305 Two pathname operators are available when expanding a variable. If the
1306 variable name begins with @samp{/} then only the last component of the
1307 pathname is substituted. For example, if @code{$@{path@}} was
1308 @samp{/foo/bar} then @code{$@{/path@}} would be expanded to @samp{bar}.
1309 Similarly, if the variable name ends with @samp{/} then all but the last
1310 component of the pathname is substituted. In the previous example,
1311 @code{$@{path/@}} would be expanded to @samp{/foo}.@refill
1313 Two domain name operators are also provided. If the variable name
1314 begins with @samp{.} then only the domain part of the name is
1315 substituted. For example, if @code{$@{rhost@}} was
1316 @samp{swan.doc.ic.ac.uk} then @code{$@{.rhost@}} would be expanded to
1317 @samp{doc.ic.ac.uk}. Similarly, if the variable name ends with @samp{.}
1318 then only the host component is substituted. In the previous example,
1319 @code{$@{rhost.@}} would be expanded to @samp{swan}.@refill
1321 Variable expansion is a two phase process. Before a location is parsed,
1322 all references to selectors, @i{eg} @code{$@{path@}}, are expanded. The
1323 location is then parsed, selections are evaluated and option assignments
1324 recorded. If there were no selections or they all succeeded the
1325 location is used and the values of the following options are expanded in
1326 the order given: @var{sublink}, @var{rfs}, @var{fs}, @var{opts},
1327 @var{remopts}, @var{mount} and @var{unmount}.
1329 Note that expansion of option values is done after @dfn{all} assignments
1330 have been completed and not in a purely left to right order as is done
1331 by the shell. This generally has the desired effect but care must be
1332 taken if one of the options references another, in which case the
1333 ordering can become significant.
1335 There are two special cases concerning variable expansion:
1339 before a map is consulted, any selectors in the name received
1340 from the kernel are expanded. For example, if the request from the
1341 kernel was for `@t{$@{arch@}}@t{.bin}' and the machine architecture
1342 was @samp{vax}, the value given to @code{$@{key@}} would be
1343 @samp{vax.bin}.@refill
1346 the value of @code{$@{rhost@}} is expanded and normalized before the
1347 other options are expanded. The normalization process strips any local
1348 sub-domain components. For example, if @code{$@{domain@}} was
1349 @samp{Berkeley.EDU} and @code{$@{rhost@}} was initially
1350 @samp{snow.Berkeley.EDU}, after the normalization it would simply be
1351 @samp{snow}. Hostname normalization is currently done in a
1352 @emph{case-dependent} manner.@refill
1355 @c======================================================================
1356 @node Selectors, Map Options, Variable Expansion, Location Format
1357 @comment node-name, next, previous, up
1358 @subsection Selectors
1361 Selectors are used to control the use of a location. It is possible to
1362 share a mount map between many machines in such a way that filesystem
1363 location, architecture and operating system differences are hidden from
1364 the users. A selector of the form @samp{arch==sun3;os==sunos4} would only
1365 apply on Sun-3s running SunOS 4.x.
1367 Selectors can be negated by using @samp{!=} instead of @samp{==}. For
1368 example to select a location on all non-Vax machines the selector
1369 @samp{arch!=vax} would be used.
1371 Selectors are evaluated left to right. If a selector fails then that
1372 location is ignored. Thus the selectors form a conjunction and the
1373 locations form a disjunction. If all the locations are ignored or
1374 otherwise fail then @i{Amd} uses the @dfn{error} filesystem
1375 (@pxref{Error Filesystem}). This is equivalent to having a location
1376 @samp{type:=error} at the end of each mount-map entry.@refill
1378 The default value of many of the selectors listed here can be overridden
1379 by an @i{Amd} command line switch or in an @i{Amd} configuration file.
1380 @xref{Amd Configuration File}.
1382 The following selectors are currently implemented.
1385 * arch Selector Variable::
1386 * autodir Selector Variable::
1387 * byte Selector Variable::
1388 * cluster Selector Variable::
1389 * domain Selector Variable::
1390 * dollar Selector Variable::
1391 * host Selector Variable::
1392 * hostd Selector Variable::
1393 * karch Selector Variable::
1394 * os Selector Variable::
1395 * osver Selector Variable::
1396 * full_os Selector Variable::
1397 * vendor Selector Variable::
1399 * key Selector Variable::
1400 * map Selector Variable::
1401 * netnumber Selector Variable::
1402 * network Selector Variable::
1403 * path Selector Variable::
1404 * wire Selector Variable::
1405 * uid Selector Variable::
1406 * gid Selector Variable::
1408 * exists Selector Function::
1409 * false Selector Function::
1410 * netgrp Selector Function::
1411 * netgrpd Selector Function::
1412 * in_network Selector Function::
1413 * true Selector Function::
1414 * xhost Selector Function::
1417 @c ----------------------------------------------------------------
1418 @node arch Selector Variable, autodir Selector Variable, Selectors, Selectors
1419 @comment node-name, next, previous, up
1420 @subsubsection arch Selector Variable
1421 @cindex arch Selector Variable
1422 @cindex arch, mount selector
1423 @cindex Mount selector; arch
1424 @cindex Selector; arch
1426 The machine architecture which was automatically determined at compile
1427 time. The architecture type can be displayed by running the command
1428 @samp{amd -v}. You can override this value also using the @code{-A}
1429 command line option. @xref{Supported Platforms}.@refill
1431 @c ----------------------------------------------------------------
1432 @node autodir Selector Variable, byte Selector Variable, arch Selector Variable, Selectors
1433 @comment node-name, next, previous, up
1434 @subsubsection autodir Selector Variable
1435 @cindex autodir Selector Variable
1436 @cindex autodir, mount selector
1437 @cindex Mount selector; autodir
1438 @cindex Selector; autodir
1440 The default directory under which to mount filesystems. This may be
1441 changed by the @code{-a} command line option. @xref{fs Option}.
1443 @c ----------------------------------------------------------------
1444 @node byte Selector Variable, cluster Selector Variable, autodir Selector Variable, Selectors
1445 @comment node-name, next, previous, up
1446 @subsubsection byte Selector Variable
1447 @cindex byte Selector Variable
1448 @cindex byte, mount selector
1449 @cindex Mount selector; byte
1450 @cindex Selector; byte
1452 The machine's byte ordering. This is either @samp{little}, indicating
1453 little-endian, or @samp{big}, indicating big-endian. One possible use
1454 is to share @samp{rwho} databases (@pxref{rwho servers}). Another is to
1455 share ndbm databases, however this use can be considered a courageous
1458 @c ----------------------------------------------------------------
1459 @node cluster Selector Variable, domain Selector Variable, byte Selector Variable, Selectors
1460 @comment node-name, next, previous, up
1461 @subsubsection cluster Selector Variable
1462 @cindex cluster Selector Variable
1463 @cindex cluster, mount selector
1464 @cindex Mount selector; cluster
1465 @cindex Selector; cluster
1467 This is provided as a hook for the name of the local cluster. This can
1468 be used to decide which servers to use for copies of replicated
1469 filesystems. @code{$@{cluster@}} defaults to the value of
1470 @code{$@{domain@}} unless a different value is set with the @code{-C}
1471 command line option.
1473 @c ----------------------------------------------------------------
1474 @node domain Selector Variable, dollar Selector Variable, cluster Selector Variable, Selectors
1475 @comment node-name, next, previous, up
1476 @subsubsection domain Selector Variable
1477 @cindex domain Selector Variable
1478 @cindex domain, mount selector
1479 @cindex Mount selector; domain
1480 @cindex Selector; domain
1482 The local domain name as specified by the @code{-d} command line option.
1483 @xref{host Selector Variable}.
1485 @c ----------------------------------------------------------------
1486 @node dollar Selector Variable, host Selector Variable, domain Selector Variable, Selectors
1487 @comment node-name, next, previous, up
1488 @subsubsection dollar Selector Variable
1489 @cindex dollar Selector Variable
1491 This is a special variable, whose sole purpose is to produce a literal
1492 dollar sign in the value of another variable. For example, if you have
1493 a remote file system whose name is @samp{/disk$s}, you can mount it by
1494 setting the remote file system variable as follows:
1497 rfs:=/disk$@{dollar@}s
1500 @c ----------------------------------------------------------------
1501 @node host Selector Variable, hostd Selector Variable, dollar Selector Variable, Selectors
1502 @comment node-name, next, previous, up
1503 @subsubsection host Selector Variable
1504 @cindex host Selector Variable
1505 @cindex host, mount selector
1506 @cindex Mount selector; host
1507 @cindex Selector; host
1509 The local hostname as determined by @b{gethostname}(2). If no domain
1510 name was specified on the command line and the hostname contains a
1511 period @samp{.} then the string before the period is used as the host
1512 name, and the string after the period is assigned to @code{$@{domain@}}.
1513 For example, if the hostname is @samp{styx.doc.ic.ac.uk} then
1514 @code{host} would be @samp{styx} and @code{domain} would be
1515 @samp{doc.ic.ac.uk}. @code{hostd} would be
1516 @samp{styx.doc.ic.ac.uk}.@refill
1518 @c ----------------------------------------------------------------
1519 @node hostd Selector Variable, karch Selector Variable, host Selector Variable, Selectors
1520 @comment node-name, next, previous, up
1521 @subsubsection hostd Selector Variable
1522 @cindex hostd Selector Variable
1523 @cindex hostd, mount selector
1524 @cindex Mount selector; hostd
1525 @cindex Selector; hostd
1527 This resolves to the @code{$@{host@}} and @code{$@{domain@}}
1528 concatenated with a @samp{.} inserted between them if required. If
1529 @code{$@{domain@}} is an empty string then @code{$@{host@}} and
1530 @code{$@{hostd@}} will be identical.
1532 @c ----------------------------------------------------------------
1533 @node karch Selector Variable, os Selector Variable, hostd Selector Variable, Selectors
1534 @comment node-name, next, previous, up
1535 @subsubsection karch Selector Variable
1536 @cindex karch Selector Variable
1537 @cindex karch, mount selector
1538 @cindex Mount selector; karch
1539 @cindex Selector; karch
1541 This is provided as a hook for the kernel architecture. This is used on
1542 SunOS 4 and SunOS 5, for example, to distinguish between different
1543 @samp{/usr/kvm} volumes. @code{$@{karch@}} defaults to the ``machine''
1544 value gotten from @b{uname}(2). If the @b{uname}(2) system call is not
1545 available, the value of @code{$@{karch@}} defaults to that of
1546 @code{$@{arch@}}. Finally, a different value can be set with the @code{-k}
1547 command line option.
1549 @c ----------------------------------------------------------------
1550 @node os Selector Variable, osver Selector Variable, karch Selector Variable, Selectors
1551 @comment node-name, next, previous, up
1552 @subsubsection os Selector Variable
1553 @cindex os Selector Variable
1554 @cindex os, mount selector
1555 @cindex Mount selector; os
1556 @cindex Selector; os
1558 The operating system. Like the machine architecture, this is
1559 automatically determined at compile time. The operating system name can
1560 be displayed by running the command @samp{amd -v}. @xref{Supported
1563 @c ----------------------------------------------------------------
1564 @node osver Selector Variable, full_os Selector Variable, os Selector Variable, Selectors
1565 @comment node-name, next, previous, up
1566 @subsubsection osver Selector Variable
1567 @cindex osver Selector Variable
1568 @cindex osver, mount selector
1569 @cindex Mount selector; osver
1570 @cindex Selector; osver
1572 The operating system version. Like the machine architecture, this is
1573 automatically determined at compile time. The operating system name can
1574 be displayed by running the command @samp{amd -v}. @xref{Supported
1577 @c ----------------------------------------------------------------
1578 @node full_os Selector Variable, vendor Selector Variable, osver Selector Variable, Selectors
1579 @comment node-name, next, previous, up
1580 @subsubsection full_os Selector Variable
1581 @cindex full_os Selector Variable
1582 @cindex full_os, mount selector
1583 @cindex Mount selector; full_os
1584 @cindex Selector; full_os
1586 The full name of the operating system, including its version. This
1587 value is automatically determined at compile time. The full operating
1588 system name and version can be displayed by running the command
1589 @samp{amd -v}. @xref{Supported Platforms}.@refill
1591 @c ----------------------------------------------------------------
1592 @node vendor Selector Variable, key Selector Variable, full_os Selector Variable, Selectors
1593 @comment node-name, next, previous, up
1594 @subsubsection vendor Selector Variable
1595 @cindex vendor Selector Variable
1596 @cindex vendor, mount selector
1597 @cindex Mount selector; vendor
1598 @cindex Selector; vendor
1600 The name of the vendor of the operating system. This value is
1601 automatically determined at compile time. The name of the vendor can be
1602 displayed by running the command @samp{amd -v}. @xref{Supported
1606 @c ----------------------------------------------------------------
1611 The following selectors are also provided. Unlike the other selectors,
1612 they vary for each lookup. Note that when the name from the kernel is
1613 expanded prior to a map lookup, these selectors are all defined as empty
1616 @c ----------------------------------------------------------------
1617 @node key Selector Variable, map Selector Variable, vendor Selector Variable, Selectors
1618 @comment node-name, next, previous, up
1619 @subsubsection key Selector Variable
1620 @cindex key Selector Variable
1621 @cindex key, mount selector
1622 @cindex Mount selector; key
1623 @cindex Selector; key
1625 The name being resolved. For example, if @file{/home} is an automount
1626 point, then accessing @file{/home/foo} would set @code{$@{key@}} to the
1627 string @samp{foo}. The key is prefixed by the @var{pref} option set in
1628 the parent mount point. The default prefix is an empty string. If the
1629 prefix was @file{blah/} then @code{$@{key@}} would be set to
1630 @file{blah/foo}.@refill
1632 @c ----------------------------------------------------------------
1633 @node map Selector Variable, netnumber Selector Variable, key Selector Variable, Selectors
1634 @comment node-name, next, previous, up
1635 @subsubsection map Selector Variable
1636 @cindex map Selector Variable
1637 @cindex map, mount selector
1638 @cindex Mount selector; map
1639 @cindex Selector; map
1641 The name of the mount map being used.
1643 @c ----------------------------------------------------------------
1644 @node netnumber Selector Variable, network Selector Variable, map Selector Variable, Selectors
1645 @comment node-name, next, previous, up
1646 @subsubsection netnumber Selector Variable
1647 @cindex netnumber Selector Variable
1648 @cindex netnumber, mount selector
1649 @cindex Mount selector; netnumber
1650 @cindex Selector; netnumber
1652 This selector is identical to the @samp{in_network} selector function,
1653 see @ref{in_network Selector Function}. It will match either the name
1654 or number of @i{any} network interface on which this host is connected
1655 to. The names and numbers of all attached interfaces are available from
1656 the output of @samp{amd -v}.
1658 @c ----------------------------------------------------------------
1659 @node network Selector Variable, path Selector Variable, netnumber Selector Variable, Selectors
1660 @comment node-name, next, previous, up
1661 @subsubsection network Selector Variable
1662 @cindex network Selector Variable
1663 @cindex network, mount selector
1664 @cindex Mount selector; network
1665 @cindex Selector; network
1667 This selector is identical to the @samp{in_network} selector function,
1668 see @ref{in_network Selector Function}. It will match either the name
1669 or number of @i{any} network interface on which this host is connected
1670 to. The names and numbers of all attached interfaces are available from
1671 the output of @samp{amd -v}.
1673 @c ----------------------------------------------------------------
1674 @node path Selector Variable, wire Selector Variable, network Selector Variable, Selectors
1675 @comment node-name, next, previous, up
1676 @subsubsection path Selector Variable
1677 @cindex path Selector Variable
1678 @cindex path, mount selector
1679 @cindex Mount selector; path
1680 @cindex Selector; path
1682 The full pathname of the name being resolved. For example
1683 @file{/home/foo} in the example above.
1685 @c ----------------------------------------------------------------
1686 @node wire Selector Variable, uid Selector Variable, path Selector Variable, Selectors
1687 @comment node-name, next, previous, up
1688 @subsubsection wire Selector Variable
1689 @cindex wire Selector Variable
1690 @cindex wire, mount selector
1691 @cindex Mount selector; wire
1692 @cindex Selector; wire
1694 This selector is identical to the @samp{in_network} selector function,
1695 see @ref{in_network Selector Function}. It will match either the name
1696 or number of @i{any} network interface on which this host is connected
1697 to. The names and numbers of all attached interfaces are available from
1698 the output of @samp{amd -v}.
1700 @c ----------------------------------------------------------------
1701 @node uid Selector Variable, gid Selector Variable, wire Selector Variable, Selectors
1702 @comment node-name, next, previous, up
1703 @subsubsection uid Selector Variable
1704 @cindex uid Selector Variable
1705 @cindex uid, mount selector
1706 @cindex Mount selector; uid
1707 @cindex Selector; uid
1709 This selector provides the numeric effective user ID (UID) of the user
1710 which last accessed an automounted path name. This simple example shows
1711 how floppy mounting can be assigned only to machine owners:
1714 floppy -type:=pcfs \
1715 uid==2301;host==shekel;dev:=/dev/floppy \
1716 uid==6712;host==titan;dev=/dev/fd0 \
1717 uid==0;dev:=/dev/fd0c \
1721 The example allows two machine owners to mount floppies on their
1722 designated workstations, allows the root user to mount on any host, and
1723 otherwise forces an error.
1725 @c ----------------------------------------------------------------
1726 @node gid Selector Variable, exists Selector Function, uid Selector Variable, Selectors
1727 @comment node-name, next, previous, up
1728 @subsubsection gid Selector Variable
1729 @cindex gid Selector Variable
1730 @cindex gid, mount selector
1731 @cindex Mount selector; gid
1732 @cindex Selector; gid
1734 This selector provides the numeric effective group ID (GID) of the user
1735 which last accessed an automounted path name.
1737 @c ----------------------------------------------------------------
1742 The following boolean functions are selectors which take an argument
1743 @i{ARG}. They return a value of true or false, and thus do not need to
1744 be compared with a value. Each of these may be negated by prepending
1745 @samp{!} to their name.
1747 @c ----------------------------------------------------------------
1748 @node exists Selector Function, false Selector Function, gid Selector Variable, Selectors
1749 @comment node-name, next, previous, up
1750 @subsubsection exists Selector Function
1751 @cindex exists Selector Function
1752 @cindex exists, boolean mount selector
1753 @cindex !exists, boolean mount selector
1754 @cindex Mount selector; exists
1755 @cindex Selector; exists
1757 If the file listed by @i{ARG} exists (via @b{lstat}(2)), this function
1758 evaluates to true. Otherwise it evaluates to false.
1760 @c ----------------------------------------------------------------
1761 @node false Selector Function, netgrp Selector Function, exists Selector Function, Selectors
1762 @comment node-name, next, previous, up
1763 @subsubsection false Selector Function
1764 @cindex false Selector Function
1765 @cindex false, boolean mount selector
1766 @cindex !false, boolean mount selector
1767 @cindex Mount selector; false
1768 @cindex Selector; false
1770 Always evaluates to false. @i{ARG} is ignored.
1772 @c ----------------------------------------------------------------
1773 @node netgrp Selector Function, netgrpd Selector Function, false Selector Function, Selectors
1774 @comment node-name, next, previous, up
1775 @subsubsection netgrp Selector Function
1776 @cindex netgrp Selector Function
1777 @cindex netgrp, boolean mount selector
1778 @cindex !netgrp, boolean mount selector
1779 @cindex Mount selector; netgrp
1780 @cindex Selector; netgrp
1782 The argument @i{ARG} of this selector is a netgroup name followed
1783 optionally by a comma and a host name. If the host name is not
1784 specified, it defaults to @code{$@{host@}}. If the host name (short
1785 name) is a member of the netgroup, this selector evaluates to true.
1786 Otherwise it evaluates to false.
1788 For example, suppose you have a netgroup @samp{ppp-hosts}, and for
1789 reasons of performance, these have a local @file{/home} partition,
1790 while all other clients on the faster network can access a shared home
1791 directory. A common map to use for both might look like the
1795 home/* netgrp(ppp-hosts);type:=link;fs:=/local/$@{key@} \
1796 !netgrp(ppp-hosts);type:=nfs;rhost:=serv1;rfs:=/remote/$@{key@}
1799 A more complex example that takes advantage of the two argument netgrp
1800 mount selector is given in the following scenario. Suppose one wants
1801 to mount the local scratch space from a each host under
1802 @file{scratch/<hostname>} and some hosts have their scratch space in a
1803 different path than others. Hosts in the netgroup @samp{apple-hosts}
1804 have their scratch space in the @file{/apple} path, where hosts in the
1805 netgroup @samp{cherry-hosts} have their scratch space in the
1806 @file{/cherry} path. For hosts that are neither in the
1807 @samp{apple-hosts} or @samp{cherry-hosts} netgroups we want to make a
1808 symlink pointing to nowhere but provide a descriptive error message in
1809 the link destination:
1812 scratch/* netgrp(apple-hosts,$@{/key@});type:=nfs;rhost:=$@{/key@};\
1814 netgrp(cherry-hosts,$@{/key@});type:=nfs;rhost:=$@{/key@};\
1816 type:=link;rfs:="no local partition for $@{/key@}"
1819 @c ----------------------------------------------------------------
1820 @node netgrpd Selector Function, in_network Selector Function, netgrp Selector Function, Selectors
1821 @comment node-name, next, previous, up
1822 @subsubsection netgrpd Selector Function
1823 @cindex netgrpd Selector Function
1824 @cindex netgrpd, boolean mount selector
1825 @cindex !netgrpd, boolean mount selector
1826 @cindex Mount selector; netgrpd
1827 @cindex Selector; netgrpd
1829 The argument @i{ARG} of this selector is a netgroup name followed
1830 optionally by a comma and a host name. If the host name is not
1831 specified, it defaults to @code{$@{hostd@}}. If the host name
1832 (fully-qualified name) is a member of the netgroup, this selector
1833 evaluates to true. Otherwise it evaluates to false.
1835 The @samp{netgrpd} function uses fully-qualified host names to match
1836 netgroup names, while the @samp{netgrp} function (@pxref{netgrp
1837 Selector Function}) uses short host names.
1839 @c ----------------------------------------------------------------
1840 @node in_network Selector Function, true Selector Function, netgrpd Selector Function, Selectors
1841 @comment node-name, next, previous, up
1842 @subsubsection in_network Selector Function
1843 @cindex in_network Selector Function
1844 @cindex in_network, boolean mount selector
1845 @cindex !in_network, boolean mount selector
1846 @cindex Mount selector; in_network
1847 @cindex Selector; in_network
1849 This selector matches against any network name or number with an
1850 optional netmask. First, if the current host has any network interface that is
1851 locally attached to the network specified in @i{ARG} (either via name or
1852 number), this selector evaluates to true.
1854 Second, @samp{in_network} supports a network/netmask syntax such as
1855 @samp{128.59.16.0/255.255.255.0}, @samp{128.59.16.0/24},
1856 @samp{128.59.16.0/0xffffff00}, or @samp{128.59.16.0/}. Using the last
1857 form, @i{Amd} will match the specified network number against the
1858 default netmasks of each of the locally attached interfaces.
1860 If the selector does not match, it evaluates to false.
1862 For example, suppose you have two servers that have an exportable
1863 @file{/opt} that smaller clients can NFS mount. The two servers are
1864 say, @samp{serv1} on network @samp{foo-net.site.com} and @samp{serv2} on
1865 network @samp{123.4.5.0}. You can write a map to be used by all clients
1866 that will attempt to mount the closest one as follows:
1869 opt in_network(foo-net.site.com);rhost:=serv1;rfs:=/opt \
1870 in_network(123.4.5.0);rhost:=serv2;rfs:=/opt \
1871 rhost:=fallback-server
1874 @c ----------------------------------------------------------------
1875 @node true Selector Function, xhost Selector Function, in_network Selector Function, Selectors
1876 @comment node-name, next, previous, up
1877 @subsubsection true Selector Function
1878 @cindex true Selector Function
1879 @cindex true, boolean mount selector
1880 @cindex !true, boolean mount selector
1881 @cindex Mount selector; true
1882 @cindex Selector; true
1884 Always evaluates to true. @i{ARG} is ignored.
1886 @c ----------------------------------------------------------------
1887 @node xhost Selector Function, , true Selector Function, Selectors
1888 @comment node-name, next, previous, up
1889 @subsubsection xhost Selector Function
1890 @cindex xhost Selector Function
1891 @cindex xhost, boolean mount selector
1892 @cindex !xhost, boolean mount selector
1893 @cindex Mount selector; xhost
1894 @cindex Selector; xhost
1897 This function compares @i{ARG} against the current hostname, similarly
1898 to the @ref{host Selector Variable}. However, this function will
1899 also match if @i{ARG} is a CNAME (DNS Canonical Name, or alias) for
1900 the current host's name.
1902 @c ================================================================
1903 @node Map Options, , Selectors, Location Format
1904 @comment node-name, next, previous, up
1905 @subsection Map Options
1907 @cindex Setting map options
1909 Options are parsed concurrently with selectors. The difference is that
1910 when an option is seen the string following the @samp{:=} is
1911 recorded for later use. As a minimum the @var{type} option must be
1912 specified. Each filesystem type has other options which must also be
1913 specified. @xref{Filesystem Types}, for details on the filesystem
1914 specific options.@refill
1916 Superfluous option specifications are ignored and are not reported
1919 The following options apply to more than one filesystem type.
1931 @node addopts Option, delay Option, Map Options, Map Options
1932 @comment node-name, next, previous, up
1933 @subsubsection addopts Option
1934 @cindex Setting additional options on a mount location
1935 @cindex Overriding or adding options to a mount
1936 @cindex addopts, mount option
1937 @cindex Mount option; addopts
1939 This option adds additional options to default options normally
1940 specified in the @samp{/defaults} entry or the defaults of the key entry
1941 being processed (@pxref{opts Option}). Normally when you specify
1942 @samp{opts} in both the @samp{/defaults} and the map entry, the latter
1943 overrides the former completely. But with @samp{addopts} it will append
1944 the options and override any conflicting ones.
1946 @samp{addopts} also overrides the value of the @samp{remopts} option
1947 (@pxref{remopts Option}), which unless specified defaults to the value
1950 Options which start with @samp{no} will override those with the same
1951 name that do not start with @samp{no} and vice verse. Special handling
1952 is given to inverted options such as @samp{soft} and @samp{hard},
1953 @samp{bg} and @samp{fg}, @samp{ro} and @samp{rw}, etc.
1955 For example, if the default options specified were
1957 opts:=rw,nosuid,intr,rsize=1024,wsize=1024,quota,posix
1960 and the ones specified in a map entry were
1963 addopts:=grpid,suid,ro,rsize=2048,quota,nointr
1966 then the actual options used would be
1969 wsize=1024,posix,grpid,suid,ro,rsize=2048,quota,nointr
1972 @node delay Option, fs Option, addopts Option, Map Options
1973 @comment node-name, next, previous, up
1974 @subsubsection delay Option
1975 @cindex Setting a delay on a mount location
1976 @cindex Delaying mounts from specific locations
1977 @cindex Primary server
1978 @cindex Secondary server
1979 @cindex delay, mount option
1980 @cindex Mount option; delay
1982 The delay, in seconds, before an attempt will be made to mount from the
1983 current location. Auxiliary data, such as network address, file handles
1984 and so on are computed regardless of this value.
1986 A delay can be used to implement the notion of primary and secondary
1987 file servers. The secondary servers would have a delay of a few
1988 seconds, thus giving the primary servers a chance to respond first.
1990 @node fs Option, opts Option, delay Option, Map Options
1991 @comment node-name, next, previous, up
1992 @subsubsection fs Option
1993 @cindex Setting the local mount point
1994 @cindex Overriding the default mount point
1995 @cindex fs, mount option
1996 @cindex Mount option; fs
1998 The local mount point. The semantics of this option vary between
2001 For NFS and UFS filesystems the value of @code{$@{fs@}} is used as the
2002 local mount point. For other filesystem types it has other meanings
2003 which are described in the section describing the respective filesystem
2004 type. It is important that this string uniquely identifies the
2005 filesystem being mounted. To satisfy this requirement, it should
2006 contain the name of the host on which the filesystem is resident and the
2007 pathname of the filesystem on the local or remote host.
2009 The reason for requiring the hostname is clear if replicated filesystems
2010 are considered. If a fileserver goes down and a replacement filesystem
2011 is mounted then the @dfn{local} mount point @dfn{must} be different from
2012 that of the filesystem which is hung. Some encoding of the filesystem
2013 name is required if more than one filesystem is to be mounted from any
2016 If the hostname is first in the path then all mounts from a particular
2017 host will be gathered below a single directory. If that server goes
2018 down then the hung mount points are less likely to be accidentally
2019 referenced, for example when @b{getcwd}(3) traverses the namespace to
2020 find the pathname of the current directory.
2022 The @samp{fs} option defaults to
2023 @code{$@{autodir@}/$@{rhost@}$@{rfs@}}. In addition,
2024 @samp{rhost} defaults to the local host name (@code{$@{host@}}) and
2025 @samp{rfs} defaults to the value of @code{$@{path@}}, which is the full
2026 path of the requested file; @samp{/home/foo} in the example above
2027 (@pxref{Selectors}). @code{$@{autodir@}} defaults to @samp{/a} but may
2028 be changed with the @code{-a} command line option. Sun's automounter
2029 defaults to @samp{/tmp_mnt}. Note that there is no @samp{/} between
2030 the @code{$@{rhost@}} and @code{$@{rfs@}} since @code{$@{rfs@}} begins
2031 with a @samp{/}.@refill
2033 @node opts Option, remopts Option, fs Option, Map Options
2034 @comment node-name, next, previous, up
2035 @subsubsection opts Option
2036 @cindex Setting system mount options
2037 @cindex Passing parameters to the mount system call
2038 @cindex mount system call
2039 @cindex mount system call flags
2040 @cindex The mount system call
2041 @cindex opts, mount option
2042 @cindex Mount option; opts
2044 The options to pass to the mount system call. A leading @samp{-} is
2045 silently ignored. The mount options supported generally correspond to
2046 those used by @b{mount}(8) and are listed below. Some additional
2047 pseudo-options are interpreted by @i{Amd} and are also listed.
2049 Unless specifically overridden, each of the system default mount options
2050 applies. Any options not recognized are ignored. If no options list is
2051 supplied the string @samp{rw,defaults} is used and all the system
2052 default mount options apply. Options which are not applicable for a
2053 particular operating system are silently ignored. For example, only 4.4BSD
2054 is known to implement the @code{compress} and @code{spongy} options.
2058 @item acdirmax=@var{n}
2059 @cindex Mount flags; acdirmax
2060 Set the maximum directory attribute cache timeout to @var{n}.
2062 @item acdirmin=@var{n}
2063 @cindex Mount flags; acdirmin
2064 Set the minimum directory attribute cache timeout to @var{n}.
2066 @item acregmax=@var{n}
2067 @cindex Mount flags; acregmax
2068 Set the maximum file attribute cache timeout to @var{n}.
2070 @item acregmin=@var{n}
2071 @cindex Mount flags; acregmin
2072 Set the minimum file attribute cache timeout to @var{n}.
2074 @item actimeo=@var{n}
2075 @cindex Mount flags; actimeo
2076 Set the overall attribute cache timeout to @var{n}.
2079 @cindex Mount flags; auto
2081 @cindex Mount flags; ignore
2082 Ignore this mount by @b{df}(1).
2085 @cindex Mount flags; cache
2086 Allow data to be cached from a remote server for this mount.
2089 @cindex Mount flags; closesession
2090 For UDF mounts, close the session when unmounting.
2093 @cindex Mount flags; compress
2094 Use NFS compression protocol.
2097 @cindex Mount flags; defperm
2098 Ignore the permission mode bits, and default file permissions to 0555,
2099 UID 0, and GID 0. Useful for CD-ROMs formatted as ISO-9660.
2102 @cindex Mount flags; dev
2103 Allow local special devices on this filesystem.
2105 @item dirmask=@var{n}
2106 @cindex Mount flags; dirmask
2107 For PCFS mounts, specify the maximum file permissions for directories
2108 in the file system. See the @samp{mask} option's description for more
2109 details. The mask value of @var{n} can be specified in decimal,
2110 octal, or hexadecimal.
2113 @cindex Mount flags; dumbtimr
2114 Turn off the dynamic retransmit timeout estimator. This may be useful
2115 for UDP mounts that exhibit high retry rates, since it is possible that
2116 the dynamically estimated timeout interval is too short.
2119 @cindex Mount flags; extatt
2120 Enable extended attributes in ISO-9660 file systems.
2123 @cindex Mount flags; fsid
2124 Set ID of filesystem.
2127 @cindex Mount flags; gens
2128 Enable generations in ISO-9660 file systems. Generations allow you to
2129 see all versions of a given file.
2131 @item gmtoff=@var{n}
2132 @cindex Mount flags; gmtoff
2133 For UDF mounts, set the time zone offset from UTC to @var{n} seconds,
2134 with positive values indicating east of the Prime Meridian. If not
2135 set, the user's current time zone will be used.
2138 @cindex Mount flags; group
2139 For PCFS and UDF mounts, set the group of the files in the file system
2140 to @var{n} (which can either be a group name or a GID number). The
2141 default group is the group of the directory on which the file system
2145 @cindex Mount flags; grpid
2146 Use BSD directory group-id semantics.
2149 @cindex Mount flags; int
2151 @cindex Mount flags; intr
2152 Allow keyboard interrupts on hard mounts.
2155 @cindex Mount flags; lock
2156 Use the NFS locking protocol (default)
2159 @cindex Mount Flags; longname
2160 For PCFS mounts, force Win95 long names.
2163 @cindex Mount flags; mask
2164 For PCFS mounts, specify the maximum file permissions for files in the
2165 file system. For example, a mask of 755 specifies that, by default,
2166 the owner should have read, write, and execute permissions for files,
2167 but others should only have read and execute permissions. Only the
2168 nine low-order bits of mask are used. The default mask is taken from
2169 the directory on which the file system is being mounted. The mask
2170 value of @var{n} can be specified in decimal, octal, or hexadecimal.
2173 @cindex Mount flags; multi
2174 Perform multi-component lookup on files.
2177 @cindex Mount flags; maxgroups
2178 Set the maximum number of groups to allow for this mount.
2181 @cindex Mount flags; nfsv3
2182 Use NFS Version 3 for this mount.
2185 @cindex Mount flags; noac
2186 Turn off the attribute cache.
2189 @cindex Mount flags; noauto
2190 This option is used by the mount command in @samp{/etc/fstab} or
2191 @samp{/etc/vfstab} and means not to mount this file system when mount -a
2195 @cindex Mount flags; nocache
2196 Do not allow data to be cached from a remote server for this
2200 @cindex Mount flags; nocasetrans
2201 Don't do case translation. Useful for CD-ROMS formatted as
2205 @cindex Mount flags; noconn
2206 Don't make a connection on datagram transports.
2209 @cindex Mount flags; nocto
2210 No close-to-open consistency.
2213 @cindex Mount flags; nodefperm
2214 Do not ignore the permission mode bits. Useful for CD-ROMS formatted as
2218 @cindex Mount flags; nodev
2220 @cindex Mount flags; nodevs
2221 Don't allow local special devices on this filesystem.
2224 @cindex Mount flags; noexec
2225 Don't allow program execution.
2228 @cindex Mount flags; noint
2229 Do not allow keyboard interrupts for this mount
2232 @cindex Mount flags; nojoliet
2233 Turn off the Joliet extensions. Useful for CD-ROMS formatted as ISO-9660.
2236 @cindex Mount flags; nolock
2237 Do not use the NFS locking protocol
2240 @cindex Mount flags; nomnttab
2241 This option is used internally to tell Amd that a Solaris 8 system using
2245 @cindex Mount flags; norrip
2246 Turn off using of the Rock Ridge Interchange Protocol (RRIP) extensions
2250 @cindex Mount flags; nosub
2251 Disallow mounts beneath this mount.
2254 @cindex Mount flags; nosuid
2255 Don't allow set-uid or set-gid executables on this filesystem.
2258 @cindex Mount flags; noversion
2259 Strip the extension @samp{;#} from the version string of files recorded
2260 on an ISO-9660 CD-ROM.
2263 @cindex Mount Flags; nowin95
2264 For PCFS mounts, completely ignore Win95 entries.
2267 @cindex Mount flags; optionstr
2268 Under Solaris 8, provide the kernel a string of options to parse and
2269 show as part of the special in-kernel mount file system.
2272 @cindex Mount flags; overlay
2273 Overlay this mount on top of an existing mount, if any.
2275 @item pgthresh=@var{n}
2276 @cindex Mount flags; pgthresh
2277 Set the paging threshold to @var{n} kilobytes.
2280 @cindex Mount flags; port
2281 Set the NFS port to @var{n}.
2284 @cindex Mount flags; posix
2285 Turn on POSIX static pathconf for mounts.
2288 @cindex Mount flags; private
2289 Use local locking instead of the NLM protocol, useful for IRIX 6 only.
2292 @cindex Mount flags; proplist
2293 Support property lists (ACLs) for this mount, useful primarily for Tru64
2297 @cindex Mount flags; proto
2298 Use transport protocol @var{s} for NFS (can be @code{"tcp"} or @code{"udp"}).
2301 @cindex Mount flags; quota
2302 Enable quota checking on this mount.
2305 @cindex Mount flags; rdonly
2307 @cindex Mount flags; ro
2308 Mount this filesystem readonly.
2311 @cindex Mount flags; resvport
2312 Use a reserved port (smaller than 1024) for remote NFS mounts. Most
2313 systems assume that, but some allow for mounts to occur on non-reserved
2314 ports. This causes problems when such a system tries to NFS mount one
2315 that requires reserved ports. It is recommended that this option always
2319 @cindex Mount flags; retrans
2320 The number of NFS retransmits made before a user error is generated by a
2321 @samp{soft} mounted filesystem, and before a @samp{hard} mounted
2322 filesystem reports @samp{NFS server @dfn{yoyo} not responding still
2326 @cindex Mount flags; retry
2327 Set the NFS retry counter.
2330 @cindex Mount flags; rrcaseins
2331 Enable the Rock Ridge Interchange Protocol (RRIP) case insensitive extensions.
2332 Useful for CD-ROMS formatted as ISO-9660.
2335 @cindex Mount flags; rrip
2336 Uses the Rock Ridge Interchange Protocol (RRIP) extensions to ISO-9660.
2339 @cindex Mount flags; rsize
2340 The NFS read packet size. You may need to set this if you are using
2341 NFS/UDP through a gateway or a slow link.
2344 @cindex Mount flags; rw
2345 Allow reads and writes on this filesystem.
2347 @item sessionnr=@var{n}
2348 @cindex Mount Flags; sessionnr
2349 For multisession UDF mounts, use session number @var{n} when mounting.
2352 @cindex Mount Flags; longname
2353 For PCFS mounts, force old DOS short names only.
2356 @cindex Mount flags; soft
2357 Give up after @dfn{retrans} retransmissions.
2360 @cindex Mount flags; spongy
2361 Like @samp{soft} for status requests, and @samp{hard} for data transfers.
2364 @cindex Mount flags; suid
2365 Allow set-uid programs on this mount.
2368 @cindex Mount flags; symttl
2369 Turn off the symbolic link cache time-to-live.
2372 @cindex Mount flags; sync
2373 Perform synchronous filesystem operations on this mount.
2376 @cindex Mount flags; tcp
2377 Use TCP/IP instead of UDP/IP, ignored if the NFS implementation does not
2378 support TCP/IP mounts.
2381 @cindex Mount flags; timeo
2382 The NFS timeout, in tenth-seconds, before a request is retransmitted.
2385 @cindex Mount flags; user
2386 For PCFS and UDF mounts, set the owner of the files in the file system
2387 to @var{n} (which can either be a user name or a UID number). The
2388 default owner is the owner of the directory on which the file system
2392 @cindex Mount flags; vers
2393 Use NFS protocol version number @var{n} (can be 2 or 3).
2396 @cindex Mount flags; wsize
2397 The NFS write packet size. You may need to set this if you are using
2398 NFS/UDP through a gateway or a slow link.
2402 The following options are implemented by @i{Amd}, rather than being
2403 passed to the kernel.
2408 @cindex Mount flags; nounmount
2409 Configures the mount so that its time-to-live will never expire. This
2410 is the default for non-network based filesystem types (such as
2411 mounting local disks, floppies, and CD-ROMs). See also the related
2414 @c Implementation broken:
2417 @cindex Mount flags; ping
2418 The interval, in seconds, between keep-alive pings. When four
2419 consecutive pings have failed the mount point is marked as hung. This
2420 interval defaults to 30 seconds; if the ping interval is set to zero,
2421 @i{Amd} will use the default 30-second interval. If the interval is
2422 set to -1 (or any other negative value), no pings are sent and the
2423 host is assumed to be always up, which can cause unmounts to hang See
2424 the @i{softlookup} option for a better alternative. Turning pings off
2425 can be useful in NFS-HA (High-Availability) sites where the NFS
2426 service rarely goes down. Setting the ping value to a large value can
2427 reduce the amount of NFS_NULL chatter on your network considerably,
2428 especially in large sites.
2430 Note that if you have multiple @i{Amd} entries using the same file
2431 server, and each entry sets a different value of N, then each time Amd
2432 mounts a new entry, the ping value will be re-evaluated (and updated,
2433 turned off, or turned back on as needed). Finally, note that NFS_NULL
2434 pings are sent for both UDP and TCP mounts, because even a hung TCP
2435 mount can cause user processes to hang.
2438 @cindex Mount flags; public
2439 Use WebNFS multi-component lookup on the public file handle instead of
2440 the mount protocol to obtain NFS file handles, as documented in the
2441 WebNFS Client Specification, RFC 2054. This means that @i{Amd} will not
2442 attempt to contact the remote portmapper or remote mountd daemon, and
2443 will only connect to the well-known NFS port 2049 or the port specified
2444 with the @i{port} mount option, thus making it easier to use NFS through
2448 @cindex Mount flags; retry=@var{n}
2449 The number of times to retry the mount system call.
2452 @cindex Mount flags; softlookup
2453 Configures @i{Amd}'s behavior with respect to already-mounted shares from
2454 NFS fileservers that are unreachable. If softlookup is specified,
2455 trying to access such a share will result in an error (EIO, which is
2456 changed from the ENOENT 6.0 used to return). If it is not specified, a
2457 regular symlink is provided and the access will probably hang
2458 in the NFS filesystem.
2460 The default behavior depends on whether the mount is 'soft' or 'hard';
2461 softlookup can be used to change this default. This is changed from 6.0
2462 which always behaved as if softlookup was specified.
2465 @cindex Mount flags; unmount
2466 Configures the mount so that its time-to-live will indeed expire (and
2467 thus may be automatically unmounted). This is also the default for
2468 network-based filesystem types (e.g., NFS). This option is useful for
2469 removable local media such as CD-ROMs, USB drives, etc. so they can
2470 expire when not in use, and get unmounted (such drives can get work
2471 out when they keep spinning). See also the related @i{nounmount}
2474 @item utimeout=@var{n}
2475 @cindex Mount flags; utimeout=@var{n}
2476 The interval, in seconds, that looked up and mounted map entries are
2477 cached. After that period of time, @i{Amd} will attempt to unmount
2478 the entries. If, however, the unmount fails (with EBUSY), then
2479 @i{Amd} will extend the mount's time-to-live by the @i{utimeout} value
2480 before the next unmount attempt is made. In fact the interval is
2481 extended before the unmount is attempted, to avoid thrashing. The
2482 default value is 120 seconds (two minutes) or as set by the @code{-w}
2483 command line option.
2486 @cindex Mount flags; xlatecookie
2487 Translate directory cookies between 32-long and 64-long lengths.
2491 @node remopts Option, sublink Option, opts Option, Map Options
2492 @comment node-name, next, previous, up
2493 @subsubsection remopts Option
2494 @cindex Setting system mount options for non-local networks
2495 @cindex remopts, mount option
2496 @cindex Mount option; remopts
2498 This option has the same use as @code{$@{opts@}} but applies only when
2499 the remote host is on a non-local network. For example, when using NFS
2500 across a gateway it is often necessary to use smaller values for the
2501 data read and write sizes. This can simply be done by specifying the
2502 small values in @var{remopts}. When a non-local host is accessed, the
2503 smaller sizes will automatically be used.
2505 @i{Amd} determines whether a host is local by examining the network
2506 interface configuration at startup. Any interface changes made after
2507 @i{Amd} has been started will not be noticed. The likely effect will
2508 be that a host may incorrectly be declared non-local.
2510 Unless otherwise set, the value of @code{$@{remopts@}} is the same as
2511 the value of @code{$@{opts@}}.
2513 @node sublink Option, type Option, remopts Option, Map Options
2514 @comment node-name, next, previous, up
2515 @subsubsection sublink Option
2516 @cindex Setting the sublink option
2517 @cindex sublink, mount option
2518 @cindex Mount option; sublink
2520 The subdirectory within the mounted filesystem to which the reference
2521 should point. This can be used to prevent duplicate mounts in cases
2522 where multiple directories in the same mounted filesystem are used.
2524 @node type Option, , sublink Option, Map Options
2525 @comment node-name, next, previous, up
2526 @subsubsection type Option
2527 @cindex Setting the filesystem type option
2528 @cindex type, mount option
2529 @cindex Mount option; type
2531 The filesystem type to be used. @xref{Filesystem Types}, for a full
2532 description of each type.@refill
2534 @c ################################################################
2535 @node Amd Command Line Options, Filesystem Types, Mount Maps, Top
2536 @comment node-name, next, previous, up
2537 @chapter @i{Amd} Command Line Options
2538 @cindex Command line options, Amd
2539 @cindex Amd command line options
2540 @cindex Overriding defaults on the command line
2542 Many of @i{Amd}'s parameters can be set from the command line. The
2543 command line is also used to specify automount points and maps.
2545 The general format of a command line is
2548 amd [@i{options}] [@{ @i{directory} @i{map-name} [-@i{map-options}] @} ...]
2551 For each directory and map-name given or specified in the
2552 @file{amd.conf} file, @i{Amd} establishes an automount point. The
2553 @dfn{map-options} may be any sequence of options or
2554 selectors---@pxref{Location Format}. The @dfn{map-options} apply only
2555 to @i{Amd}'s mount point.
2557 @samp{type:=toplvl;cache:=mapdefault;fs:=$@{map@}} is the default value for the
2558 map options. Default options for a map are read from a special entry in
2559 the map whose key is the string @samp{/defaults}. When default options
2560 are given they are prepended to any options specified in the mount-map
2561 locations as explained in @ref{Map Defaults}.
2563 The @dfn{options} are any combination of those listed below.
2565 Once the command line has been parsed, the automount points are mounted.
2566 The mount points are created if they do not already exist, in which case they
2567 will be removed when @i{Amd} exits.
2568 Finally, @i{Amd} disassociates itself from its controlling terminal and
2569 forks into the background.
2571 Note: Even if @i{Amd} has been built with @samp{-DDEBUG} (via
2572 @code{configure --enable-debug}), it will still background itself and
2573 disassociate itself from the controlling terminal. To use a debugger it
2574 is necessary to specify @samp{-D daemon} on the command line.
2575 However, even with all of this, mounts and unmounts are performed in the
2576 background, and @i{Amd} will always fork before doing them. Therefore,
2577 debugging what happens closely during un/mounts is more challenging.
2579 @emph{All} of @i{Amd}'s command options (save @code{-F} and @code{-T})
2580 can be specified in the @file{amd.conf} file. @xref{Amd Configuration
2581 File}. If @i{Amd} is invoked without any command line options, it will
2582 default to using the configuration file @file{/etc/amd.conf}, if one
2586 * -a Option:: Automount directory.
2587 * -c Option:: Cache timeout interval.
2588 * -d Option:: Domain name.
2589 * -k Option:: Kernel architecture.
2590 * -l Option:: Log file.
2591 * -n Option:: Hostname normalization.
2592 * -o Option:: Operating system version.
2593 * -p Option:: Output process id.
2594 * -r Option:: Restart existing mounts.
2595 * -t Option:: Kernel RPC timeout.
2596 * -v Option:: Version information.
2597 * -w Option:: Wait interval after failed unmount.
2598 * -x Option:: Log options.
2599 * -y Option:: NIS domain.
2600 * -A Option:: Operating system Architecture.
2601 * -C Option:: Cluster name.
2602 * -D Option:: Debug flags.
2603 * -F Option:: Amd configuration file.
2604 * -H Option:: Show brief help.
2605 * -O Option:: Operating system name.
2606 * -S Option:: Lock executable pages in memory.
2607 * -T Option:: Set tag for configuration file.
2610 @c ----------------------------------------------------------------
2611 @node -a Option, -c Option, Amd Command Line Options, Amd Command Line Options
2612 @comment node-name, next, previous, up
2613 @section @code{-a} @var{directory}
2614 @cindex Automount directory
2615 @cindex Setting the default mount directory
2617 Specifies the default mount directory. This option changes the variable
2618 @code{$@{autodir@}} which otherwise defaults to @file{/a}. For example,
2619 some sites prefer @file{/amd} or @file{/n}.
2625 @c ----------------------------------------------------------------
2626 @node -c Option, -d Option, -a Option, Amd Command Line Options
2627 @comment node-name, next, previous, up
2628 @section @code{-c} @var{cache-interval}
2629 @cindex Cache interval
2630 @cindex Interval before a filesystem times out
2631 @cindex Setting the interval before a filesystem times out
2632 @cindex Changing the interval before a filesystem times out
2634 Selects the period, in seconds, for which a name is cached by @i{Amd}.
2635 If no reference is made to the volume in this period, @i{Amd} discards
2636 the volume name to filesystem mapping.
2638 Once the last reference to a filesystem has been removed, @i{Amd}
2639 attempts to unmount the filesystem. If the unmount fails the interval
2640 is extended by a further period as specified by the @samp{-w} command
2641 line option or by the @samp{utimeout} mount option.
2643 The default @dfn{cache-interval} is 300 seconds (five minutes).
2645 @c ----------------------------------------------------------------
2646 @node -d Option, -k Option, -c Option, Amd Command Line Options
2647 @comment node-name, next, previous, up
2648 @section @code{-d} @var{domain}
2650 @cindex Setting the local domain name
2651 @cindex Overriding the local domain name
2653 Specifies the host's domain. This sets the internal variable
2654 @code{$@{domain@}} and affects the @code{$@{hostd@}} variable.
2656 If this option is not specified and the hostname already contains the
2657 local domain then that is used, otherwise the default value of
2658 @code{$@{domain@}} is @samp{unknown.domain}.
2660 For example, if the local domain was @samp{doc.ic.ac.uk}, @i{Amd} could
2661 be started as follows:
2664 amd -d doc.ic.ac.uk ...
2667 @c ----------------------------------------------------------------
2668 @node -k Option, -l Option, -d Option, Amd Command Line Options
2669 @comment node-name, next, previous, up
2670 @section @code{-k} @var{kernel-architecture}
2671 @cindex Setting the Kernel architecture
2673 Specifies the kernel architecture of the system. This is usually the
2674 output of @samp{uname -m} (the ``machine'' value gotten from
2675 @b{uname}(2)). If the @b{uname}(2) system call is not available, the
2676 value of @code{$@{karch@}} defaults to that of @code{$@{arch@}}.
2678 The only effect of this option is to set the variable @code{$@{karch@}}.
2680 This option would be used as follows:
2683 amd -k `arch -k` ...
2686 @c ----------------------------------------------------------------
2687 @node -l Option, -n Option, -k Option, Amd Command Line Options
2688 @comment node-name, next, previous, up
2689 @section @code{-l} @var{log-option}
2690 @cindex Log filename
2691 @cindex Setting the log file
2692 @cindex Using syslog to log errors
2695 Selects the form of logging to be made. Several special @dfn{log-options}
2700 If @dfn{log-option} is the string @samp{syslog}, @i{Amd} will use the
2701 @b{syslog}(3) mechanism. If your system supports syslog facilities, then
2702 the default facility used is @samp{LOG_DAEMON}.
2705 @cindex syslog facility; specifying an alternate
2706 When using syslog, if you wish to change the facility, append its name
2707 to the log option name, delimited by a single colon. For example, if
2708 @dfn{log-options} is the string @samp{syslog:local7} then @b{Amd} will
2709 log messages via @b{syslog}(3) using the @samp{LOG_LOCAL7} facility. If
2710 the facility name specified is not recognized, @i{Amd} will default to
2711 @samp{LOG_DAEMON}. Note: while you can use any syslog facility
2712 available on your system, it is generally a bad idea to use those
2713 reserved for other services such as @samp{kern}, @samp{lpr},
2717 If @dfn{log-option} is the string @samp{/dev/stderr}, @i{Amd} will use
2718 standard error, which is also the default target for log messages. To
2719 implement this, @i{Amd} simulates the effect of the @samp{/dev/fd}
2723 Any other string is taken as a filename to use for logging. Log
2724 messages are appended to the file if it already exists, otherwise a new
2725 file is created. The file is opened once and then held open, rather
2726 than being re-opened for each message.
2728 Normally, when long-running daemons hold an open file descriptor on a
2729 log file, it is impossible to ``rotate'' the log file and compress older
2730 logs on a daily basis. The daemon needs to be told to discard (via
2731 @b{close}(2)) its file handle, and re-open the log file. This is done
2732 using @code{amq -l} @i{log-option}. @xref{Amq -l option}.
2734 If the @samp{syslog} option is specified but the system does not support
2735 syslog or if the named file cannot be opened or created, @i{Amd} will
2736 use standard error. Error messages generated before @i{Amd} has
2737 finished parsing the command line are printed on standard error.
2739 Since @i{Amd} tends to generate a lot of logging information (especially
2740 if debugging was turned on), and due to it being an important program
2741 running on the system, it is usually best to log to a separate disk
2742 file. In that case @i{Amd} would be started as follows:
2745 amd -l /var/log/amd ...
2748 @c ----------------------------------------------------------------
2749 @node -n Option, -o Option, -l Option, Amd Command Line Options
2750 @comment node-name, next, previous, up
2752 @cindex Hostname normalization
2753 @cindex Aliased hostnames
2754 @cindex Resolving aliased hostnames
2755 @cindex Normalizing hostnames
2757 Normalizes the remote hostname before using it. Normalization is done
2758 by replacing the value of @code{$@{rhost@}} with the (generally fully
2759 qualified) primary name returned by a hostname lookup.
2761 This option should be used if several names are used to refer to a
2762 single host in a mount map.
2764 @c ----------------------------------------------------------------
2765 @node -o Option, -p Option, -n Option, Amd Command Line Options
2766 @comment node-name, next, previous, up
2767 @section @code{-o} @var{op-sys-ver}
2768 @cindex Operating System version
2769 @cindex Setting the Operating System version
2771 Overrides the compiled-in version number of the operating system, with
2772 @var{op-sys-ver}. Useful when the built-in version is not desired for
2773 backward compatibility reasons. For example, if the built-in version is
2774 @samp{2.5.1}, you can override it to @samp{5.5.1}, and use older maps
2775 that were written with the latter in mind.
2777 @c ----------------------------------------------------------------
2778 @node -p Option, -r Option, -o Option, Amd Command Line Options
2779 @comment node-name, next, previous, up
2782 @cindex Displaying the process id
2783 @cindex process id of Amd daemon
2784 @cindex pid file, creating with -p option
2785 @cindex Creating a pid file
2787 Causes @i{Amd}'s process id to be printed on standard output.
2788 This can be redirected to a suitable file for use with kill:
2791 amd -p > /var/run/amd.pid ...
2794 This option only has an affect if @i{Amd} is running in daemon mode.
2795 If @i{Amd} is started with the @code{-D daemon} debug flag, this
2798 @c ----------------------------------------------------------------
2799 @node -r Option, -t Option, -p Option, Amd Command Line Options
2800 @comment node-name, next, previous, up
2802 @cindex Restarting existing mounts
2803 @cindex Picking up existing mounts
2805 Tells @i{Amd} to restart existing mounts (@pxref{Inheritance Filesystem}).
2806 @c @dfn{This option will be made the default in the next release.}
2808 @c ----------------------------------------------------------------
2809 @node -t Option, -v Option, -r Option, Amd Command Line Options
2810 @comment node-name, next, previous, up
2811 @section @code{-t} @var{timeout.retransmit}
2812 @cindex Setting Amd's RPC parameters
2814 Specifies the RPC @dfn{timeout} interval and the @dfn{retransmit}
2815 counter used by the kernel to communicate to @i{Amd}. These are used to
2816 set the @samp{timeo} and @samp{retrans} mount options, respectively.
2817 The default timeout is 0.8 seconds, and the default number of
2818 retransmissions is 11.
2820 @i{Amd} relies on the kernel RPC retransmit mechanism to trigger mount
2821 retries. The values of these parameters change the overall retry
2822 interval. Too long an interval gives poor interactive response; too
2823 short an interval causes excessive retries.
2825 @c ----------------------------------------------------------------
2826 @node -v Option, -w Option, -t Option, Amd Command Line Options
2827 @comment node-name, next, previous, up
2829 @cindex Version information
2830 @cindex Discovering version information
2831 @cindex How to discover your version of Amd
2833 Print version information on standard error and then exit. The output
2837 Copyright (c) 1997-1999 Erez Zadok
2838 Copyright (c) 1990 Jan-Simon Pendry
2839 Copyright (c) 1990 Imperial College of Science, Technology & Medicine
2840 Copyright (c) 1990 The Regents of the University of California.
2841 am-utils version 6.0a15 (build 61).
2842 Built by ezk@@example.com on date Wed Oct 22 15:21:03 EDT 1997.
2843 cpu=sparc (big-endian), arch=sun4, karch=sun4u.
2844 full_os=solaris2.5.1, os=sos5, osver=5.5.1, vendor=sun.
2845 Map support for: root, passwd, union, nisplus, nis, ndbm, file, error.
2846 AMFS: nfs, link, nfsx, nfsl, host, linkx, program, union, inherit,
2847 ufs, lofs, hsfs, pcfs, auto, direct, toplvl, error.
2848 FS: autofs, cachefs, cdfs, lofs, nfs, nfs3, pcfs, tfs, tmpfs, udf, ufs.
2849 Network 1: wire="mcl-lab-net.cs.columbia.edu" (netnumber=128.59.13).
2850 Network 2: wire="14-net.cs.columbia.edu" (netnumber=128.59.14).
2851 Network 3: wire="old-net.cs.columbia.edu" (netnumber=128.59.16).
2854 The information includes the version number, number of times @i{Amd} was
2855 compiled on the local system, release date and name of the release.
2856 Following come the cpu type, byte ordering, and the architecture and
2857 kernel architecture as @code{$@{arch@}} and @code{$@{karch@}},
2858 respectively. The next line lists the operating system full name, short
2859 name, version, and vendor. These four values correspond to the
2860 variables @code{$@{full_os@}}, @code{$@{os@}}, @code{$@{osver@}}, and
2861 @code{$@{vendor@}}, respectively. @xref{Supported Platforms}.
2863 Then come a list of map types supported, filesystems internally
2864 supported by @i{Amd} (AMFS), and generic filesystems available (FS).
2865 Finally all known networks (if any) of this host are listed by name
2866 and number. They are available via the variables
2867 @code{$@{wire@}} or @code{$@{network@}}, and
2868 @code{$@{netnumber@}} (@pxref{Selectors}) or the @samp{in_network}
2869 selector function (@pxref{in_network Selector Function}).
2871 @c ----------------------------------------------------------------
2872 @node -w Option, -x Option, -v Option, Amd Command Line Options
2873 @comment node-name, next, previous, up
2874 @section @code{-w} @var{wait-timeout}
2875 @cindex Setting the interval between unmount attempts
2876 @cindex unmount attempt backoff interval
2878 Selects the interval in seconds between unmount attempts after the
2879 initial time-to-live has expired.
2881 This defaults to 120 seconds (two minutes).
2883 @c ----------------------------------------------------------------
2884 @node -x Option, -y Option, -w Option, Amd Command Line Options
2885 @comment node-name, next, previous, up
2886 @section @code{-x} @var{opts}
2887 @cindex Log message selection
2888 @cindex Selecting specific log messages
2889 @cindex How to select log messages
2890 @cindex syslog priorities
2892 Specifies the type and verbosity of log messages. @dfn{opts} is
2893 a comma separated list selected from the following options:
2897 Fatal errors (cannot be turned off)
2899 Non-fatal errors (cannot be turned off)
2901 Non-fatal user errors
2905 Alias for @code{warn}
2907 Information messages
2911 Additional statistics
2915 An alias for "fatal,error,user,warning,info".
2918 Initially a set of default logging flags is enabled. This is as if
2921 @samp{-x fatal,error,user,warning,info}
2922 had been selected. The command line is
2923 parsed and logging is controlled by the @code{-x} option. The very first
2924 set of logging flags is saved and can not be subsequently disabled using
2925 @i{Amq}. This default set of options is useful for general production
2928 The @samp{info} messages include details of what is mounted and
2929 unmounted and when filesystems have timed out. If you want to have the
2930 default set of messages without the @samp{info} messages then you simply
2931 need @samp{-x noinfo}. The messages given by @samp{user} relate to
2932 errors in the mount maps, so these are useful when new maps are
2933 installed. The following table lists the syslog priorities used for each
2934 of the message types.@refill
2955 The options can be prefixed by the string @samp{no} to indicate
2956 that this option should be turned off. For example, to obtain all
2957 but @samp{info} messages the option @samp{-x all,noinfo} would be used.
2959 If @i{Amd} was built with debugging enabled the @code{debug} option is
2960 automatically enabled regardless of the command line options.
2962 @c ----------------------------------------------------------------
2963 @node -y Option, -A Option, -x Option, Amd Command Line Options
2964 @comment node-name, next, previous, up
2965 @section @code{-y} @var{NIS-domain}
2966 @cindex NIS (YP) domain name
2967 @cindex Overriding the NIS (YP) domain name
2968 @cindex Setting the NIS (YP) domain name
2969 @cindex YP domain name
2971 Selects an alternate NIS domain. This is useful for debugging and
2972 cross-domain shared mounting. If this flag is specified, @i{Amd}
2973 immediately attempts to bind to a server for this domain.
2974 @c @i{Amd} refers to NIS maps when it starts, unless the @code{-m} option
2975 @c is specified, and whenever required in a mount map.
2977 @c ----------------------------------------------------------------
2978 @node -A Option, -C Option, -y Option, Amd Command Line Options
2979 @comment node-name, next, previous, up
2980 @section @code{-A} @var{architecture}
2981 @cindex Setting the operating system architecture
2983 Specifies the OS architecture of the system.
2984 The only effect of this option is to set the variable @code{$@{arch@}}.
2986 This option would be used as follows:
2992 @c ----------------------------------------------------------------
2993 @node -C Option, -D Option, -A Option, Amd Command Line Options
2994 @comment node-name, next, previous, up
2995 @section @code{-C} @var{cluster-name}
2996 @cindex Cluster names
2997 @cindex Setting the cluster name
2999 Specifies the name of the cluster of which the local machine is a member.
3000 The only effect is to set the variable @code{$@{cluster@}}.
3001 The @dfn{cluster-name} is will usually obtained by running another command which uses
3002 a database to map the local hostname into a cluster name.
3003 @code{$@{cluster@}} can then be used as a selector to restrict mounting of
3005 If this option is not given, @code{$@{cluster@}} has the same value as @code{$@{domain@}}.
3006 This would be used as follows:
3009 amd -C `clustername` ...
3012 @c ----------------------------------------------------------------
3013 @node -D Option, -F Option, -C Option, Amd Command Line Options
3014 @comment node-name, next, previous, up
3015 @section @code{-D} @var{opts}
3016 @cindex Debug options
3017 @cindex Setting debug flags
3019 Controls the verbosity and coverage of the debugging trace; @dfn{opts}
3020 is a comma separated list of debugging options. The @code{-D} option is
3021 only available if @i{Amd} was compiled with @samp{-DDEBUG}, or
3022 configured with @code{configure --enable-debug}. The memory debugging
3023 facilities (@samp{mem}) are only available if @i{Amd} was compiled with
3024 @samp{-DDEBUG_MEM} (in addition to @samp{-DDEBUG}), or configured with
3025 @code{configure --enable-debug=mem}.
3027 The most common options to use are @samp{-D trace} and @samp{-D test}
3028 (which turns on all the useful debug options). As usual, every option
3029 can be prefixed with @samp{no} to turn it off.
3033 all options (excluding hrtime and mtab)
3035 ``sensible'' default options (all--excluding hrtime, mtab, and xdrtrace)
3037 full debug options plus mtab,nodaemon,nofork,noamq
3039 register @i{Amd} with the RPC portmapper, for @i{Amq}
3043 fork child worker (hlfsd only)
3047 print high resolution time stamps (only if @b{syslog}(3) is not used)
3049 @cindex debugging hesiod resolver service
3050 @cindex Hesiod; turning on RES_DEBUG
3051 info service specific debugging (hesiod, nis, etc.) In the case of
3052 hesiod maps, turns on the hesiod RES_DEBUG internal debugging option.
3054 trace memory allocations. Needs to be explicitly enabled at compile
3055 time with --enable-debug=mem.
3057 use local mount-table file (defaults to @file{/tmp/mtab}, @pxref{debug_mtab_file Parameter})
3059 show readdir progress
3061 debug string munging
3063 trace RPC protocol and NFS mount arguments
3068 You may also refer to the program source for a more detailed explanation
3069 of the available options.
3071 @c ----------------------------------------------------------------
3072 @node -F Option, -H Option, -D Option, Amd Command Line Options
3073 @comment node-name, next, previous, up
3074 @section @code{-F} @var{conf-file}
3075 @cindex Amd configuration file; specifying name
3076 @cindex Amd configuration file
3077 @cindex amd.conf file
3079 Specify an @i{Amd} configuration file @var{conf-file} to use. For a
3080 description of the format and syntax, @pxref{Amd Configuration File}.
3081 This configuration file is used to specify any options in lieu of typing
3082 many of them on the command line. The @file{amd.conf} file includes
3083 directives for every command line option @i{Amd} has, and many more that
3084 are only available via the configuration file facility. The
3085 configuration file specified by this option is processed after all other
3086 options had been processed, regardless of the actual location of this
3087 option on the command line.
3089 @c ----------------------------------------------------------------
3090 @node -H Option, -O Option, -F Option, Amd Command Line Options
3091 @comment node-name, next, previous, up
3093 @cindex Displaying brief help
3094 @cindex Help; showing from Amd
3096 Print a brief help and usage string.
3098 @c ----------------------------------------------------------------
3099 @node -O Option, -S Option, -H Option, Amd Command Line Options
3100 @comment node-name, next, previous, up
3101 @section @code{-O} @var{op-sys-name}
3102 @cindex Operating System name
3103 @cindex Setting the Operating System name
3105 Overrides the compiled-in name of the operating system, with
3106 @var{op-sys-name}. Useful when the built-in name is not desired for
3107 backward compatibility reasons. For example, if the build in name is
3108 @samp{sunos5}, you can override it to the old name @samp{sos5}, and use
3109 older maps which were written with the latter in mind.
3111 @c ----------------------------------------------------------------
3112 @node -S Option, -T Option, -O Option, Amd Command Line Options
3113 @comment node-name, next, previous, up
3115 @cindex plock; using
3116 @cindex mlockall; using
3117 @cindex locking executable pages in memory
3119 Do @emph{not} lock the running executable pages of @i{Amd} into memory.
3120 To improve @i{Amd}'s performance, systems that support the @b{plock}(3)
3122 call lock the @i{Amd} process into memory. This way there is less
3123 chance the operating system will schedule, page out, and swap the
3124 @i{Amd} process as needed. This tends to improve @i{Amd}'s performance,
3125 at the cost of reserving the memory used by the @i{Amd} process (making
3126 it unavailable for other processes). If this behavior is not desired,
3127 use the @code{-S} option.
3129 @c ----------------------------------------------------------------
3130 @node -T Option, , -S Option, Amd Command Line Options
3131 @comment node-name, next, previous, up
3132 @section @code{-T} @var{tag}
3133 @cindex Tags for Amd configuration file
3134 @cindex Configuration file; tags
3136 Specify a tag to use with @file{amd.conf}. All map entries tagged with
3137 @var{tag} will be processed. Map entries that are not tagged are always
3138 processed. Map entries that are tagged with a tag other than @var{tag}
3139 will not be processed.
3141 @c ################################################################
3142 @node Filesystem Types, Amd Configuration File, Amd Command Line Options, Top
3143 @comment node-name, next, previous, up
3144 @chapter Filesystem Types
3145 @cindex Filesystem types
3147 @cindex Types of filesystem
3149 To mount a volume, @i{Amd} must be told the type of filesystem to be
3150 used. Each filesystem type typically requires additional information
3151 such as the fileserver name for NFS.
3153 From the point of view of @i{Amd}, a @dfn{filesystem} is anything that
3154 can resolve an incoming name lookup. An important feature is support
3155 for multiple filesystem types. Some of these filesystems are
3156 implemented in the local kernel and some on remote fileservers, whilst
3157 the others are implemented internally by @i{Amd}.@refill
3159 The two common filesystem types are UFS and NFS. Four other user
3160 accessible filesystems (@samp{link}, @samp{program}, @samp{auto} and
3161 @samp{direct}) are also implemented internally by @i{Amd} and these are
3162 described below. There are two additional filesystem types internal to
3163 @i{Amd} which are not directly accessible to the user (@samp{inherit}
3164 and @samp{error}). Their use is described since they may still have an
3165 effect visible to the user.@refill
3168 * Network Filesystem:: A single NFS filesystem.
3169 * Network Host Filesystem:: NFS mount a host's entire export tree.
3170 * Network Filesystem Group:: An atomic group of NFS filesystems.
3171 * Unix Filesystem:: Native disk filesystem.
3172 * Caching Filesystem:: Caching from remote server filesystem.
3173 * CD-ROM Filesystem:: ISO9660 CD ROM.
3174 * UDF Filesystem:: Universal Disk Format filesystem.
3175 * Loopback Filesystem:: Local loopback-mount filesystem.
3176 * Memory/RAM Filesystem:: A memory or RAM-based filesystem.
3177 * Null Filesystem:: 4.4BSD's loopback-mount filesystem.
3178 * Floppy Filesystem:: MS-DOS Floppy filesystem.
3179 * Translucent Filesystem:: The directory merging filesystem.
3180 * Shared Memory+Swap Filesystem:: Sun's tmpfs filesystem.
3181 * User ID Mapping Filesystem:: 4.4BSD's umapfs filesystem.
3182 * Program Filesystem:: Generic Program mounts.
3183 * Symbolic Link Filesystem:: Local link.
3184 * Symbolic Link Filesystem II:: Local link referencing existing filesystem.
3185 * NFS-Link Filesystem:: Link if path exists, NFS otherwise.
3186 * Automount Filesystem::
3187 * Direct Automount Filesystem::
3188 * Union Filesystem::
3189 * Error Filesystem::
3190 * Top-level Filesystem::
3192 * Inheritance Filesystem::
3195 @c ----------------------------------------------------------------
3196 @node Network Filesystem, Network Host Filesystem, Filesystem Types, Filesystem Types
3197 @comment node-name, next, previous, up
3198 @section Network Filesystem (@samp{nfs})
3200 @cindex Mounting an NFS filesystem
3201 @cindex How to mount and NFS filesystem
3202 @cindex nfs, filesystem type
3203 @cindex Filesystem type; nfs
3205 The @dfn{nfs} (@samp{type:=nfs}) filesystem type provides access to Sun's NFS.
3208 The following options must be specified:
3211 @cindex rhost, mount option
3212 @cindex Mount option; rhost
3214 the remote fileserver. This must be an entry in the hosts database. IP
3215 addresses are not accepted. The default value is taken
3216 from the local host name (@code{$@{host@}}) if no other value is
3219 @cindex rfs, mount option
3220 @cindex Mount option; rfs
3222 the remote filesystem.
3223 If no value is specified for this option, an internal default of
3224 @code{$@{path@}} is used.
3227 NFS mounts require a two stage process. First, the @dfn{file handle} of
3228 the remote file system must be obtained from the server. Then a mount
3229 system call must be done on the local system. @i{Amd} keeps a cache
3230 of file handles for remote file systems. The cache entries have a
3231 lifetime of a few minutes.
3233 If a required file handle is not in the cache, @i{Amd} sends a request
3234 to the remote server to obtain it.
3235 @c @i{Amd} @dfn{does not} wait for
3236 @c a response; it notes that one of the locations needs retrying, but
3237 @c continues with any remaining locations. When the file handle becomes
3238 @c available, and assuming none of the other locations was successfully
3239 @c mounted, @i{Amd} will retry the mount. This mechanism allows several
3240 @c NFS filesystems to be mounted in parallel.
3241 @c @footnote{The mechanism
3242 @c is general, however NFS is the only filesystem
3243 @c for which the required hooks have been written.}
3244 @c The first one which responds with a valid file handle will be used.
3246 Historically, this documentation has maintained that @i{Amd} will try
3247 all the locations in parallel and use the first one which responds
3248 with a valid file handle. This has not been the case for quite some
3249 time, however. Instead, @i{Amd} will go through each location, one by
3250 one, and will only skip to the next one if the previous one either
3254 An NFS entry might be:
3257 jsp host!=charm;type:=nfs;rhost:=charm;rfs:=/home/charm;sublink:=jsp
3260 The mount system call and any unmount attempts are always done
3261 in a new task to avoid the possibility of blocking @i{Amd}.
3263 @c ----------------------------------------------------------------
3264 @node Network Host Filesystem, Network Filesystem Group, Network Filesystem, Filesystem Types
3265 @comment node-name, next, previous, up
3266 @section Network Host Filesystem (@samp{host})
3267 @cindex Network host filesystem
3268 @cindex Mounting entire export trees
3269 @cindex How to mount all NFS exported filesystems
3270 @cindex host, filesystem type
3271 @cindex Filesystem type; host
3273 @c NOTE: the current implementation of the @dfn{host} filesystem type
3274 @c sometimes fails to maintain a consistent view of the remote mount tree.
3275 @c This happens when the mount times out and only some of the remote mounts
3276 @c are successfully unmounted. To prevent this from occurring, use the
3277 @c @samp{nounmount} mount option.
3279 The @dfn{host} (@samp{type:=host}) filesystem allows access to the entire export tree of an
3280 NFS server. The implementation is layered above the @samp{nfs}
3281 implementation so keep-alives work in the same way. The only option
3282 which needs to be specified is @samp{rhost} which is the name of the
3283 fileserver to mount.
3285 The @samp{host} filesystem type works by querying the mount daemon on
3286 the given fileserver to obtain its export list. @i{Amd} then obtains
3287 filehandles for each of the exported filesystems. Any errors at this
3288 stage cause that particular filesystem to be ignored. Finally each
3289 filesystem is mounted. Again, errors are logged but ignored. One
3290 common reason for mounts to fail is that the mount point does not exist.
3291 Although @i{Amd} attempts to automatically create the mount point, it
3292 may be on a remote filesystem to which @i{Amd} does not have write
3295 When an attempt to unmount a @samp{host} filesystem mount fails, @i{Amd}
3296 remounts any filesystems which had successfully been unmounted. To do
3297 this @i{Amd} queries the mount daemon again and obtains a fresh copy of
3298 the export list. @i{Amd} then tries to mount any exported filesystems
3299 which are not currently mounted.
3301 Sun's automounter provides a special @samp{-hosts} map. To achieve the
3302 same effect with @i{Amd} requires two steps. First a mount map must
3303 be created as follows:
3306 * type:=host;rhost:=$@{key@};fs:=$@{autodir@}/$@{rhost@}/root
3310 and then start @i{Amd} with the following command
3317 where @samp{net.map} is the name of map described above. Note that the
3318 value of @code{$@{fs@}} is overridden in the map. This is done to avoid
3319 a clash between the mount tree and any other filesystem already mounted
3320 from the same fileserver.
3322 If different mount options are needed for different hosts then
3323 additional entries can be added to the map, for example
3326 host2 opts:=ro,nosuid,soft
3330 would soft mount @samp{host2} read-only.
3332 @c ----------------------------------------------------------------
3333 @node Network Filesystem Group, Unix Filesystem, Network Host Filesystem, Filesystem Types
3334 @comment node-name, next, previous, up
3335 @section Network Filesystem Group (@samp{nfsx})
3336 @cindex Network filesystem group
3337 @cindex Atomic NFS mounts
3338 @cindex Mounting an atomic group of NFS filesystems
3339 @cindex How to mount an atomic group of NFS filesystems
3340 @cindex nfsx, filesystem type
3341 @cindex Filesystem type; nfsx
3343 The @dfn{nfsx} (@samp{type:=nfsx}) filesystem allows a group of filesystems to be mounted
3344 from a single NFS server. The implementation is layered above the
3345 @samp{nfs} implementation so keep-alives work in the same way.
3347 @emph{WARNING}: @samp{nfsx} is meant to be a ``last resort'' kind of
3348 solution. It is racy and poorly supported. The authors @emph{highly}
3349 recommend that other solutions be considered before relying on it.
3351 The options are the same as for the @samp{nfs} filesystem with one
3352 difference for @samp{rfs}, as explained below.
3355 The following options should be specified:
3359 the remote fileserver. The default value is taken from the local
3360 host name (@code{$@{host@}}) if no other value is specified.
3363 is a list of filesystems to mount, and must be specified.
3364 The list is in the form of a comma separated strings.
3371 pub type:=nfsx;rhost:=gould;\
3372 rfs:=/public,/,graphics,usenet;fs:=$@{autodir@}/$@{rhost@}/root
3375 The first string defines the root of the tree, and is applied as a
3376 prefix to the remaining members of the list which define the individual
3377 filesystems. The first string is @emph{not} used as a filesystem name.
3378 A serial operation is used to determine the local mount points to
3379 ensure a consistent layout of a tree of mounts.
3381 Here, the @emph{three} filesystems, @samp{/public},
3382 @samp{/public/graphics} and @samp{/public/usenet}, would be mounted.@refill
3384 A local mount point, @code{$@{fs@}}, @emph{must} be specified. The
3385 default local mount point will not work correctly in the general case.
3386 A suggestion is to use @samp{fs:=$@{autodir@}/$@{rhost@}/root}.@refill
3388 @c ----------------------------------------------------------------
3389 @node Unix Filesystem, Caching Filesystem, Network Filesystem Group, Filesystem Types
3390 @comment node-name, next, previous, up
3391 @section Unix Filesystem (@samp{ufs}, @samp{xfs}, or @samp{efs})
3392 @cindex Unix filesystem
3396 @cindex Mounting a UFS filesystem
3397 @cindex Mounting a local disk
3398 @cindex How to mount a UFS filesystems
3399 @cindex How to mount a local disk
3400 @cindex Disk filesystems
3401 @cindex ufs, filesystem type
3402 @cindex Filesystem type; ufs
3403 @cindex xfs, filesystem type
3404 @cindex Filesystem type; xfs
3405 @cindex efs, filesystem type
3406 @cindex Filesystem type; efs
3408 The @dfn{ufs} (@samp{type:=ufs}) filesystem type provides access to the system's standard
3409 disk filesystem---usually a derivative of the Berkeley Fast Filesystem.
3412 The following option must be specified:
3415 @cindex dev, mount option
3416 @cindex Mount option; dev
3418 the block special device to be mounted.
3421 A UFS entry might be:
3424 jsp host==charm;type:=ufs;dev:=/dev/sd0d;sublink:=jsp
3427 UFS is the default Unix disk-based file system, which Am-utils picks up
3428 during the autoconfiguration phase. Some systems have more than one
3429 type, such as IRIX, that comes with EFS (Extent File System) and XFS
3430 (Extended File System). In those cases, you may explicitly set the file
3431 system type, by using entries such:
3434 ez1 type:=efs;dev:=/dev/xd0a
3435 ez2 type:=xfs;dev:=/dev/sd3c
3438 The UFS/XFS/EFS filesystems are never timed out by default, i.e. they
3439 will never be unmounted by @i{Amd}. If automatic unmounting is
3440 desired, the ``unmount'' option should be added to the mount options
3443 @c ----------------------------------------------------------------
3444 @node Caching Filesystem, CD-ROM Filesystem, Unix Filesystem, Filesystem Types
3445 @comment node-name, next, previous, up
3446 @section Caching Filesystem (@samp{cachefs})
3447 @cindex Caching Filesystem
3448 @cindex cachefs, filesystem type
3449 @cindex Filesystem type; cachefs
3451 The @dfn{cachefs} (@samp{type:=cachefs}) filesystem caches files from
3452 one location onto another, presumably providing faster access. It is
3453 particularly useful to cache from a larger and remote (slower) NFS
3454 partition to a smaller and local (faster) UFS directory.
3457 The following options must be specified:
3460 @cindex cachedir, mount option
3461 @cindex Mount option; cachedir
3463 the directory where the cache is stored.
3465 the path name to the ``back file system'' to be cached from.
3467 the ``front file system'' mount point to the cached files, where @i{Amd}
3468 will set a symbolic link pointing to.
3471 A CacheFS entry for, say, the @file{/import} @i{Amd} mount point, might
3475 copt type:=cachefs;cachedir:=/cache;rfs:=/import/opt;fs:=/n/import/copt
3478 Access to the pathname @file{/import/copt} will follow a symbolic link
3479 to @file{/n/import/copt}. The latter is the mount point for a caching
3480 file system, that caches from @file{/import/opt} to @file{/cache}.
3482 The cachefs filesystem is never timed out by default, i.e. it will
3483 never be unmounted by @i{Amd}. If automatic unmounting is desired, the
3484 ``unmount'' option should be added to the mount options for the entry.
3488 @item This file system is currently only implemented for Solaris 2.x!
3489 @item Before being used for the first time, the cache directory @i{must} be
3490 initialized with @samp{cfsadmin -c @var{cachedir}}. See the manual page for
3491 @b{cfsadmin}(1M) for more information.
3492 @item The ``back file system'' mounted must be a complete file system, not
3493 a subdirectory thereof; otherwise you will get an error ``Invalid Argument''.
3494 @item If @i{Amd} aborts abnormally, the state of the cache may be
3495 inconsistent, requiring running the command @file{fsck -F cachefs
3496 @var{cachedir}}. Otherwise you will get the error ``No Space Left on Device''.
3499 @c ----------------------------------------------------------------
3500 @node CD-ROM Filesystem, UDF Filesystem, Caching Filesystem, Filesystem Types
3501 @comment node-name, next, previous, up
3502 @section CD-ROM Filesystem (@samp{cdfs})
3503 @cindex CD-ROM Filesystem
3504 @cindex cdfs, filesystem type
3505 @cindex Filesystem type; cdfs
3507 The @dfn{cdfs} (@samp{type:=cdfs}) filesystem mounts a CD-ROM with an
3508 ISO9660 format filesystem on it.
3511 The following option must be specified:
3514 @cindex dev, mount option
3515 @cindex Mount option; dev
3517 the block special device to be mounted.
3520 Some operating systems will fail to mount read-only CDs unless the
3521 @samp{ro} option is specified. A cdfs entry might be:
3524 cdfs os==sunos4;type:=cdfs;dev:=/dev/sr0 \
3525 os==sunos5;addopts:=ro;type:=cdfs;dev:=/dev/dsk/c0t6d0s2
3528 @c ----------------------------------------------------------------
3529 @node UDF Filesystem, Loopback Filesystem, CD-ROM Filesystem, Filesystem Types
3530 @comment node-name, next, previous, up
3531 @section CD-ROM Filesystem (@samp{udf})
3532 @cindex CD-ROM Filesystem
3533 @cindex udf, filesystem type
3534 @cindex Filesystem type; udf
3536 The @dfn{udf} (@samp{type:=udf}) filesystem mounts media with a
3537 Universal Disk Format (UDF) filesystem on it, e.g., a video DVD.
3540 The following option must be specified:
3543 @cindex dev, mount option
3544 @cindex Mount option; dev
3546 the block special device to be mounted.
3549 Some operating systems will fail to mount read-only media unless the
3550 @samp{ro} option is specified. A udf entry might be:
3553 udf os==sunos4;type:=udf;dev:=/dev/sr0 \
3554 os==sunos5;addopts:=ro;type:=udf;dev:=/dev/dsk/c0t6d0s2
3557 @c ----------------------------------------------------------------
3558 @node Loopback Filesystem, Memory/RAM Filesystem, UDF Filesystem, Filesystem Types
3559 @comment node-name, next, previous, up
3560 @section Loopback Filesystem (@samp{lofs})
3561 @cindex Loopback Filesystem
3562 @cindex lofs, filesystem type
3563 @cindex Filesystem type; lofs
3565 The @dfn{lofs} (@samp{type:=lofs}) filesystem is also called the
3566 loopback filesystem. It mounts a local directory on another, thus
3567 providing mount-time binding to another location (unlike symbolic
3570 The loopback filesystem is particularly useful within the context of a
3571 chroot-ed directory (via @b{chroot}(2)), to provide access to
3572 directories otherwise inaccessible.
3575 The following option must be specified:
3578 @cindex rfs, mount option
3579 @cindex Mount option; rfs
3581 the pathname to be mounted on top of @code{$@{fs@}}.
3584 Usually, the FTP server runs in a chroot-ed environment, for security
3585 reasons. In this example, lofs is used to provide a subdirectory within
3586 a user's home directory, also available for public ftp.
3589 lofs type:=lofs;rfs:=/home/ezk/myftpdir;fs:=/usr/ftp/pub/ezk
3592 @c ----------------------------------------------------------------
3593 @node Memory/RAM Filesystem, Null Filesystem, Loopback Filesystem, Filesystem Types
3594 @comment node-name, next, previous, up
3595 @section Memory/RAM Filesystem (@samp{mfs})
3596 @cindex Memory/RAM Filesystem
3597 @cindex mfs, filesystem type
3598 @cindex Filesystem type; mfs
3600 The @dfn{mfs} (@samp{type:=mfs}) filesystem is available in 4.4BSD,
3601 Linux, and other systems. It creates a filesystem in a portion of the
3602 system's memory, thus providing very fast file (volatile) access.
3604 XXX: THIS FILESYSTEM IS NOT IMPLEMENTED YET!
3606 @c ----------------------------------------------------------------
3607 @node Null Filesystem, Floppy Filesystem, Memory/RAM Filesystem, Filesystem Types
3608 @comment node-name, next, previous, up
3609 @section Null Filesystem (@samp{nullfs})
3610 @cindex Null Filesystem
3611 @cindex nullfs, filesystem type
3612 @cindex Filesystem type; nullfs
3614 The @dfn{nullfs} (@samp{type:=nullfs}) filesystem is available from 4.4BSD,
3615 and is very similar to the loopback filesystem, @dfn{lofs}.
3617 XXX: THIS FILESYSTEM IS NOT IMPLEMENTED YET!
3619 @c ----------------------------------------------------------------
3620 @node Floppy Filesystem, Translucent Filesystem, Null Filesystem, Filesystem Types
3621 @comment node-name, next, previous, up
3622 @section Floppy Filesystem (@samp{pcfs})
3623 @cindex Floppy Filesystem
3624 @cindex pcfs, filesystem type
3625 @cindex Filesystem type; pcfs
3627 The @dfn{pcfs} (@samp{type:=pcfs}) filesystem mounts a floppy previously
3628 formatted for the MS-DOS format.
3631 The following option must be specified:
3634 @cindex dev, mount option
3635 @cindex Mount option; dev
3637 the block special device to be mounted.
3640 A pcfs entry might be:
3643 pcfs os==sunos4;type:=pcfs;dev:=/dev/fd0 \
3644 os==sunos5;type:=pcfs;dev:=/dev/diskette
3647 @c ----------------------------------------------------------------
3648 @node Translucent Filesystem, Shared Memory+Swap Filesystem, Floppy Filesystem, Filesystem Types
3649 @comment node-name, next, previous, up
3650 @section Translucent Filesystem (@samp{tfs})
3651 @cindex Translucent Filesystem
3652 @cindex tfs, filesystem type
3653 @cindex Filesystem type; tfs
3655 The @dfn{tfs} (@samp{type:=tfs}) filesystem is an older version of the
3656 4.4BSD @dfn{unionfs}.
3658 XXX: THIS FILESYSTEM IS NOT IMPLEMENTED YET!
3660 @c ----------------------------------------------------------------
3661 @node Shared Memory+Swap Filesystem, User ID Mapping Filesystem, Translucent Filesystem, Filesystem Types
3662 @comment node-name, next, previous, up
3663 @section Shared Memory+Swap Filesystem (@samp{tmpfs})
3664 @cindex Shared Memory and Swap Filesystem
3665 @cindex tmpfs, filesystem type
3666 @cindex Filesystem type; tmpfs
3668 The @dfn{tmpfs} (@samp{type:=tmpfs}) filesystem shares memory between a
3669 the swap device and the rest of the system. It is generally used to
3670 provide a fast access @file{/tmp} directory, one that uses memory that
3671 is otherwise unused. This filesystem is available in SunOS 4.x and 5.x.
3673 XXX: THIS FILESYSTEM IS NOT IMPLEMENTED YET!
3675 @c ----------------------------------------------------------------
3676 @node User ID Mapping Filesystem, Program Filesystem, Shared Memory+Swap Filesystem, Filesystem Types
3677 @comment node-name, next, previous, up
3678 @section User ID Mapping Filesystem (@samp{umapfs})
3679 @cindex User ID Mapping Filesystem
3680 @cindex umapfs, filesystem type
3681 @cindex Filesystem type; umapfs
3683 The @dfn{umapfs} (@samp{type:=umapfs}) filesystem maps User IDs of file
3684 ownership, and is available from 4.4BSD.
3686 XXX: THIS FILESYSTEM IS NOT IMPLEMENTED YET!
3688 @c ----------------------------------------------------------------
3689 @node Program Filesystem, Symbolic Link Filesystem, User ID Mapping Filesystem, Filesystem Types
3690 @comment node-name, next, previous, up
3691 @section Program Filesystem (@samp{program})
3692 @cindex Program filesystem
3693 @cindex Mount a filesystem under program control
3694 @cindex program, filesystem type
3695 @cindex Filesystem type; program
3697 The @dfn{program} (@samp{type:=program}) filesystem type allows a
3698 program to be run whenever a mount or unmount is required. This allows
3699 easy addition of support for other filesystem types, such as MIT's
3700 Remote Virtual Disk (RVD) which has a programmatic interface via the
3701 commands @samp{rvdmount} and @samp{rvdunmount}.
3704 Both of the following options must be specified:
3707 @cindex mount, mount option
3708 @cindex Mount option; mount
3710 the program which will perform the mount.
3712 @cindex unmount, mount option
3713 @cindex umount, mount option
3714 @cindex Mount option; unmount
3715 @cindex Mount option; umount
3718 the program which will perform the unmount. For convenience, you may
3719 use either @samp{unmount} or @samp{umount} but not both. If neither
3720 is defined, @i{Amd} will default to @samp{umount $@{fs@}} (the actual
3721 unmount program pathname will be automatically determined at the time
3722 GNU @code{configure} runs.)
3725 The exit code from these two programs is interpreted as a Unix error
3726 code. As usual, exit code zero indicates success. To execute the
3727 program, @i{Amd} splits the string on whitespace to create an array of
3728 substrings. Single quotes @samp{'} can be used to quote whitespace
3729 if that is required in an argument. There is no way to escape or change
3730 the single quote character.
3732 To run e.g. the program @samp{rvdmount} with a host name and filesystem as
3733 arguments, it would be specified by
3734 @samp{fs:=$@{autodir@}$@{path@};type:=program;mount:="/etc/rvdmount
3735 rvdmount fserver $@{fs@}";unmount:="/etc/rdvumount rvdumount $@{fs@}"}.
3737 The first element in the array is taken as the pathname of the program
3738 to execute. The other members of the array form the argument vector
3739 to be passed to the program, @dfn{including argument zero}. The array
3740 is exactly the same as the array passed to the execv() system call
3741 (man execv for details). The split string must have at least two
3742 elements. The programs are directly executed by @i{Amd}, not via a
3743 shell. Therefore, if a script is to be used as a mount/umount
3744 program, it @dfn{must} begin with a @code{#!} interpreter specification.
3746 Often, this program mount type is used for Samba mounts, where you
3747 need a double slash in pathnames. However, @i{Amd} normalizes
3748 sequences of slashes into one slash. Therefore, you must use an
3749 escaped slash, preceded by an escaped backslash. So to get a double
3750 slash in the mount command, you need the eight character sequence
3751 @samp{\\\/\\\/} in your map. For example:
3753 @samp{mount="/sbin/mount mount -r -t smbfs -o-N,-Ihostname \\\/\\\/guest@@venus/mp3"}
3755 If a filesystem type is to be heavily used, it may be worthwhile adding
3756 a new filesystem type into @i{Amd}, but for most uses the program
3757 filesystem should suffice.
3759 When the program is run, standard input and standard error are inherited
3760 from the current values used by @i{Amd}. Standard output is a
3761 duplicate of standard error. The value specified with the @code{-l}
3762 command line option has no effect on standard error.
3764 @i{Amd} guarantees that the mountpoint will be created before calling
3765 the mount program, and that it will be removed after the umount
3766 program returns success.
3768 @c ----------------------------------------------------------------
3769 @node Symbolic Link Filesystem, Symbolic Link Filesystem II, Program Filesystem, Filesystem Types
3770 @comment node-name, next, previous, up
3771 @section Symbolic Link Filesystem (@samp{link})
3772 @cindex Symbolic link filesystem
3773 @cindex Referencing part of the local name space
3774 @cindex Mounting part of the local name space
3775 @cindex How to reference part of the local name space
3776 @cindex link, filesystem type
3777 @cindex symlink, link filesystem type
3778 @cindex Filesystem type; link
3780 Each filesystem type creates a symbolic link to point from the volume
3781 name to the physical mount point. The @samp{link} filesystem does the
3782 same without any other side effects. This allows any part of the
3783 machines name space to be accessed via @i{Amd}.
3785 One common use for the symlink filesystem is @file{/homes} which can be
3786 made to contain an entry for each user which points to their
3787 (auto-mounted) home directory. Although this may seem rather expensive,
3788 it provides a great deal of administrative flexibility.
3791 The following option must be defined:
3795 The value of @var{fs} option specifies the destination of the link, as
3796 modified by the @var{sublink} option. If @var{sublink} is non-null, it
3797 is appended to @code{$@{fs@}}@code{/} and the resulting string is used
3801 The @samp{link} filesystem can be thought of as identical to the
3802 @samp{ufs} filesystem but without actually mounting anything.
3804 An example entry might be:
3807 jsp host==charm;type:=link;fs:=/home/charm;sublink:=jsp
3809 which would return a symbolic link pointing to @file{/home/charm/jsp}.
3811 @c ----------------------------------------------------------------
3812 @node Symbolic Link Filesystem II, NFS-Link Filesystem, Symbolic Link Filesystem, Filesystem Types
3813 @comment node-name, next, previous, up
3814 @section Symbolic Link Filesystem II (@samp{linkx})
3815 @cindex Symbolic link filesystem II
3816 @cindex Referencing an existing part of the local name space
3817 @cindex Mounting an existing part of the local name space
3818 @cindex How to reference an existing part of the local name space
3819 @cindex linkx, filesystem type
3820 @cindex symlink, linkx filesystem type
3821 @cindex Filesystem type; linkx
3823 The @dfn{linkx} (@samp{type:=linkx}) filesystem type is identical to @samp{link} with the
3824 exception that the target of the link must exist. Existence is checked
3825 with the @b{lstat}(2) system call.
3827 The @samp{linkx} filesystem type is particularly useful for wildcard map
3828 entries. In this case, a list of possible targets can be given and
3829 @i{Amd} will choose the first one which exists on the local machine.
3831 @c ----------------------------------------------------------------
3832 @node NFS-Link Filesystem, Automount Filesystem, Symbolic Link Filesystem II, Filesystem Types
3833 @comment node-name, next, previous, up
3834 @section NFS-Link Filesystem (@samp{nfsl})
3835 @cindex NFS-Link filesystem II
3836 @cindex Referencing an existing part of the name space if target exists
3837 @cindex Mounting a remote part of the name space if target is missing
3838 @cindex Symlink if target exists, NFS otherwise
3839 @cindex nfsl, filesystem type
3840 @cindex symlink, nfsl filesystem type
3841 @cindex Filesystem type; nfsl
3843 The @dfn{nfsl} (@samp{type:=nfsl}) filesystem type is a combination of two others:
3844 @samp{link} and @samp{nfs}. If the local host name is equal to the
3845 value of @code{$@{rhost@}} @emph{and} the target pathname listed in
3846 @code{$@{fs@}} exists, @samp{nfsl} will behave exactly as
3847 @samp{type:=link}, and refer to the target as a symbolic link. If the
3848 local host name is not equal to the value of @code{$@{rhost@}}, or if
3849 the target of the link does not exist, @i{Amd} will treat it as
3850 @samp{type:=nfs}, and will mount a remote pathname for it.
3852 The @samp{nfsl} filesystem type is particularly useful as a shorthand
3853 for the more cumbersome and yet one of the most popular @i{Amd}
3854 entries. For example, you can simplify all map entries that look like:
3857 zing -fs:=/n/shekel/u/zing \
3858 host!=shekel;type:=nfs;rhost:=shekel;rfs:=$@{fs@} \
3859 host==shekel;type:=link
3865 zing -fs:=/n/shekel/u/zing \
3866 exists($@{fs@});type:=link \
3867 !exists($@{fs@});type:=nfs;rhost:=shekel;rfs:=$@{fs@}
3873 zing type:=nfsl;fs:=/n/shekel/u/zing;rhost:=shekel;rfs:=$@{fs@}
3876 Not just does it make the maps smaller and simpler, but it avoids
3877 possible mistakes that often happen when forgetting to set up the two
3878 entries (one for @samp{type:=nfs} and the other for @samp{type:=link})
3879 necessary to perform transparent mounts of existing or remote mounts.
3881 @c ----------------------------------------------------------------
3882 @node Automount Filesystem, Direct Automount Filesystem, NFS-Link Filesystem, Filesystem Types
3883 @comment node-name, next, previous, up
3884 @section Automount Filesystem (@samp{auto})
3885 @cindex Automount filesystem
3886 @cindex Map cache types
3887 @cindex Setting map cache parameters
3888 @cindex How to set map cache parameters
3889 @cindex How to start an indirect automount point
3890 @cindex auto, filesystem type
3891 @cindex Filesystem type; auto
3892 @cindex SIGHUP signal
3893 @cindex Map cache synchronizing
3894 @cindex Synchronizing the map cache
3895 @cindex Map cache options
3896 @cindex Regular expressions in maps
3898 The @dfn{auto} (@samp{type:=auto}) filesystem type creates a new automount point below an
3899 existing automount point. Top-level automount points appear as system
3900 mount points. An automount mount point can also appear as a
3901 sub-directory of an existing automount point. This allows some
3902 additional structure to be added, for example to mimic the mount tree of
3905 The following options may be specified:
3908 @cindex cache, mount map option
3909 @cindex Mount map option; cache
3911 specifies whether the data in this mount-map should be
3912 cached. The default value is @samp{none}, in which case
3913 no caching is done in order to conserve memory.
3915 However, better performance and reliability can be obtained by caching
3916 some or all of a mount-map.
3918 If the cache option specifies @samp{all},
3919 the entire map is enumerated when the mount point is created.
3921 If the cache option specifies @samp{inc}, caching is done incrementally
3922 as and when data is required.
3923 Some map types do not support cache mode @samp{all}, in which case @samp{inc}
3924 is used whenever @samp{all} is requested.
3926 Caching can be entirely disabled by using cache mode @samp{none}.
3928 If the cache option specifies @samp{regexp} then the entire map will be
3929 enumerated and each key will be treated as an egrep-style regular
3930 expression. The order in which a cached map is searched does not
3931 correspond to the ordering in the source map so the regular expressions
3932 should be mutually exclusive to avoid confusion.
3934 Each mount map type has a default cache type, usually @samp{inc}, which
3935 can be selected by specifying @samp{mapdefault}.
3937 The cache mode for a mount map can only be selected on the command line.
3938 Starting @i{Amd} with the command:
3941 amd /homes hesiod.homes -cache:=inc
3944 will cause @samp{/homes} to be automounted using the @dfn{Hesiod} name
3945 server with local incremental caching of all successfully resolved names.
3947 All cached data is forgotten whenever @i{Amd} receives a @samp{SIGHUP}
3948 signal and, if cache @samp{all} mode was selected, the cache will be
3949 reloaded. This can be used to inform @i{Amd} that a map has been
3950 updated. In addition, whenever a cache lookup fails and @i{Amd} needs
3951 to examine a map, the map's modify time is examined. If the cache is
3952 out of date with respect to the map then it is flushed as if a
3953 @samp{SIGHUP} had been received.
3955 An additional option (@samp{sync}) may be specified to force @i{Amd} to
3956 check the map's modify time whenever a cached entry is being used. For
3957 example, an incremental, synchronized cache would be created by the
3961 amd /homes hesiod.homes -cache:=inc,sync
3965 specifies the name of the mount map to use for the new mount point.
3967 Arguably this should have been specified with the @code{$@{rfs@}} option but
3968 we are now stuck with it due to historical accident.
3970 @c %If the string @samp{.} is used then the same map is used;
3971 @c %in addition the lookup prefix is set to the name of the mount point followed
3972 @c %by a slash @samp{/}.
3973 @c %This is the same as specifying @samp{fs:=\$@{map@};pref:=\$@{key@}/}.
3977 alters the name that is looked up in the mount map. If
3978 @code{$@{pref@}}, the @dfn{prefix}, is non-null then it is prepended
3979 to the name requested by the kernel @dfn{before} the map is
3980 searched. The default prefix is the prefix of the parent map (if any)
3981 with name of the auto node appended to it. That means if you want no
3982 prefix you must say so in the map: @samp{pref:=null}.
3985 Normally, @samp{auto} style maps are not browsable even if you turn on
3986 directory browsability (@pxref{browsable_dirs Parameter}). To enable
3987 browsing entries in @samp{auto} maps, specify @samp{opts:=browsable}
3988 or @samp{opts:=fullybrowsable} in
3989 the description of this map.
3993 The server @samp{dylan.doc.ic.ac.uk} has two user disks:
3994 @samp{/dev/dsk/2s0} and @samp{/dev/dsk/5s0}. These are accessed as
3995 @samp{/home/dylan/dk2} and @samp{/home/dylan/dk5} respectively. Since
3996 @samp{/home} is already an automount point, this naming is achieved with
3997 the following map entries:@refill
4000 dylan type:=auto;fs:=$@{map@};pref:=$@{key@}/
4001 dylan/dk2 type:=ufs;dev:=/dev/dsk/2s0
4002 dylan/dk5 type:=ufs;dev:=/dev/dsk/5s0
4005 @c ----------------------------------------------------------------
4006 @node Direct Automount Filesystem, Union Filesystem, Automount Filesystem, Filesystem Types
4007 @comment node-name, next, previous, up
4008 @section Direct Automount Filesystem (@samp{direct})
4009 @cindex Direct automount filesystem
4010 @cindex How to start a direct automount point
4011 @cindex direct, filesystem type
4012 @cindex Filesystem type; direct
4014 The @dfn{direct} (@samp{type:=direct}) filesystem is almost identical to
4015 the automount filesystem. Instead of appearing to be a directory of
4016 mount points, it appears as a symbolic link to a mounted filesystem.
4017 The mount is done at the time the link is accessed. @xref{Automount
4018 Filesystem}, for a list of required options.
4020 Direct automount points are created by specifying the @samp{direct}
4021 filesystem type on the command line:
4024 amd ... /usr/man auto.direct -type:=direct
4027 where @samp{auto.direct} would contain an entry such as:
4030 usr/man -type:=nfs;rfs:=/usr/man \
4031 rhost:=man-server1 rhost:=man-server2
4034 In this example, @samp{man-server1} and @samp{man-server2} are file
4035 servers which export copies of the manual pages. Note that the key
4036 which is looked up is the name of the automount point without the
4039 Note that the implementation of the traditional @dfn{direct} filesystem is
4040 essentially a hack (pretending that the root of an NFS filesystem is a
4041 symlink) and many modern operating systems get very unhappy about
4042 it. For example, Linux kernel 2.4+ completely disallows it, and Solaris
4043 2.8 fails to unmount it when @i{Amd} shuts down. Therefore, the use of
4044 the traditional @dfn{direct} filesystem is strongly discouraged; it is
4045 only semi-supported, at best.
4047 The autofs implementations that permit direct mounts are fully
4048 supported, however. That currently includes all versions of
4049 Solaris. Linux autofs does NOT support direct mounts at all.
4051 @c ----------------------------------------------------------------
4052 @node Union Filesystem, Error Filesystem, Direct Automount Filesystem, Filesystem Types
4053 @comment node-name, next, previous, up
4054 @section Union Filesystem (@samp{union})
4055 @cindex Union filesystem
4056 @cindex union, filesystem type
4057 @cindex Filesystem type; union
4059 The @dfn{union} (@samp{type:=union}) filesystem type allows the contents of several
4060 directories to be merged and made visible in a single directory. This
4061 can be used to overcome one of the major limitations of the Unix mount
4062 mechanism which only allows complete directories to be mounted.
4064 For example, supposing @file{/tmp} and @file{/var/tmp} were to be merged
4065 into a new directory called @file{/mtmp}, with files in @file{/var/tmp}
4066 taking precedence. The following command could be used to achieve this
4070 amd ... /mtmp union:/tmp:/var/tmp -type:=union
4073 Currently, the unioned directories must @emph{not} be automounted. That
4074 would cause a deadlock. This seriously limits the current usefulness of
4075 this filesystem type and the problem will be addressed in a future
4078 Files created in the union directory are actually created in the last
4079 named directory. This is done by creating a wildcard entry which points
4080 to the correct directory. The wildcard entry is visible if the union
4081 directory is listed, so allowing you to see which directory has
4084 The files visible in the union directory are computed at the time
4085 @i{Amd} is started, and are not kept up-to-date with respect to the
4086 underlying directories. Similarly, if a link is removed, for example
4087 with the @samp{rm} command, it will be lost forever.
4089 @c ----------------------------------------------------------------
4090 @node Error Filesystem, Top-level Filesystem, Union Filesystem, Filesystem Types
4091 @comment node-name, next, previous, up
4092 @section Error Filesystem (@samp{error})
4093 @cindex Error filesystem
4094 @cindex error, filesystem type
4095 @cindex Filesystem type; error
4097 The @dfn{error} (@samp{type:=error}) filesystem type is used internally as a catch-all in the
4098 case where none of the other filesystems was selected, or some other
4099 error occurred. Lookups and mounts always fail with ``No such file or
4100 directory''. All other operations trivially succeed.
4102 The error filesystem is not directly accessible.
4104 @c ----------------------------------------------------------------
4105 @node Top-level Filesystem, Root Filesystem, Error Filesystem, Filesystem Types
4106 @comment node-name, next, previous, up
4107 @section Top-level Filesystem (@samp{toplvl})
4108 @cindex Top level filesystem
4109 @cindex toplvl, filesystem type
4110 @cindex Filesystem type; toplvl
4112 The @dfn{toplvl} (@samp{type:=toplvl}) filesystems is derived from the @samp{auto} filesystem
4113 and is used to mount the top-level automount nodes. Requests of this
4114 type are automatically generated from the command line arguments.
4116 @c ----------------------------------------------------------------
4117 @node Root Filesystem, Inheritance Filesystem, Top-level Filesystem, Filesystem Types
4118 @comment node-name, next, previous, up
4119 @section Root Filesystem (@samp{root})
4120 @cindex Root filesystem
4121 @cindex root, filesystem type
4122 @cindex Filesystem type; root
4124 The @dfn{root} (@samp{type:=root}) filesystem type acts as an internal
4125 placeholder onto which @i{Amd} can pin @samp{toplvl} mounts. Only one
4126 node of this type need ever exist and one is created automatically
4127 during startup. The effect of having more than one root node is
4130 The root filesystem is not directly accessible.
4132 @c ----------------------------------------------------------------
4133 @node Inheritance Filesystem, , Root Filesystem, Filesystem Types
4134 @comment node-name, next, previous, up
4135 @section Inheritance Filesystem (@samp{inherit})
4136 @cindex Inheritance filesystem
4137 @cindex Nodes generated on a restart
4138 @cindex inherit, filesystem type
4139 @cindex Filesystem type; inherit
4141 The @dfn{inheritance} (@samp{type:=inherit}) filesystem is not directly
4142 accessible. Instead, internal mount nodes of this type are
4143 automatically generated when @i{Amd} is started with the @code{-r} option.
4144 At this time the system mount table is scanned to locate any filesystems
4145 which are already mounted. If any reference to these filesystems is
4146 made through @i{Amd} then instead of attempting to mount it, @i{Amd}
4147 simulates the mount and @dfn{inherits} the filesystem. This allows a
4148 new version of @i{Amd} to be installed on a live system simply by
4149 killing the old daemon with @samp{SIGTERM} and starting the new one.@refill
4151 This filesystem type is not generally visible externally, but it is
4152 possible that the output from @samp{amq -m} may list @samp{inherit} as
4153 the filesystem type. This happens when an inherit operation cannot
4154 be completed for some reason, usually because a fileserver is down.
4156 @c ################################################################
4157 @node Amd Configuration File, Run-time Administration, Filesystem Types, Top
4158 @comment node-name, next, previous, up
4159 @chapter Amd Configuration File
4160 @cindex Amd Configuration File
4163 The @samp{amd.conf} file is the configuration file for @i{Amd}, as part
4164 of the am-utils suite. This file contains runtime configuration
4165 information for the @i{Amd} automounter program.
4169 * The Global Section::
4170 * Regular Map Sections::
4171 * Common Parameters::
4172 * Global Parameters::
4173 * Regular Map Parameters::
4174 * amd.conf Examples::
4177 @c ================================================================
4178 @node File Format, The Global Section, Amd Configuration File, Amd Configuration File
4179 @comment node-name, next, previous, up
4180 @section File Format
4181 @cindex amd.conf file format
4183 The @samp{amd.conf} file consists of sections and parameters. A section
4184 begins with the name of the section in square brackets @samp{[]} and
4185 continues until the next section begins or the end of the file is reached.
4186 Sections contain parameters of the form @samp{name = value}.
4188 The file is line-based --- that is, each newline-terminated line
4189 represents either a comment, a section name or a parameter. No
4190 line-continuation syntax is available.
4192 Section names, parameter names and their values are case sensitive.
4194 Only the first equals sign in a parameter is significant. Whitespace
4195 before or after the first equals sign is discarded. Leading, trailing
4196 and internal whitespace in section and parameter names is irrelevant.
4197 Leading and trailing whitespace in a parameter value is discarded.
4198 Internal whitespace within a parameter value is not allowed, unless the
4199 whole parameter value is quoted with double quotes as in @samp{name =
4202 Any line beginning with a pound sign @samp{#} is ignored, as are lines
4203 containing only whitespace.
4205 The values following the equals sign in parameters are all either a
4206 string (no quotes needed if string does not include spaces) or a
4207 boolean, which may be given as @samp{yes}/@samp{no}. Case is significant in all
4208 values. Some items such as cache timeouts are numeric.
4210 @c ================================================================
4211 @node The Global Section, Regular Map Sections, File Format, Amd Configuration File
4212 @comment node-name, next, previous, up
4213 @section The Global Section
4214 @cindex amd.conf global section
4216 The global section must be specified as @samp{[global]}. Parameters in
4217 this section either apply to @i{Amd} as a whole, or to all other regular map
4218 sections which follow. There should be only one global section defined
4219 in one configuration file.
4221 It is highly recommended that this section be specified first in the
4222 configuration file. If it is not, then regular map sections which
4223 precede it will not use global values defined later.
4225 @c ================================================================
4226 @node Regular Map Sections, Common Parameters, The Global Section, Amd Configuration File
4227 @comment node-name, next, previous, up
4228 @section Regular Map Sections
4229 @cindex amd.conf regular map sections
4231 Parameters in regular (non-global) sections apply to a single map entry.
4232 For example, if the map section @samp{[/homes]} is defined, then all
4233 parameters following it will be applied to the @file{/homes}
4234 @i{Amd}-managed mount point.
4236 @c ================================================================
4237 @node Common Parameters, Global Parameters, Regular Map Sections, Amd Configuration File
4238 @comment node-name, next, previous, up
4239 @section Common Parameters
4240 @cindex amd.conf common parameters
4242 These parameters can be specified either in the global or a map-specific
4243 section. Entries specified in a map-specific section override the default
4244 value or one defined in the global section. If such a common parameter is
4245 specified only in the global section, it is applicable to all regular map
4246 sections that follow.
4249 * autofs_use_lofs Parameter::
4250 * browsable_dirs Parameter::
4251 * map_defaults Parameter::
4252 * map_options Parameter::
4253 * map_type Parameter::
4254 * mount_type Parameter::
4255 * search_path Parameter::
4256 * selectors_in_defaults Parameter::
4257 * sun_map_syntax Parameter::
4260 @c ----------------------------------------------------------------
4261 @node autofs_use_lofs Parameter, browsable_dirs Parameter, Common Parameters, Common Parameters
4262 @comment node-name, next, previous, up
4263 @subsection @t{autofs_use_lofs} Parameter
4264 @cindex autofs_use_lofs Parameter
4266 (type=string, default=@samp{yes}).
4267 When set to @samp{yes}, @i{Amd}'s autofs code will use lofs-type
4268 (loopback) mounts for @code{type:=link} mounts, as well as several
4269 other cases that require local references. This has the advantage
4270 that @i{Amd} does not use a secondary mount point and users do not see
4271 external pathnames (the infamous @code{/bin/pwd} problem, where it
4272 reports a different path than the user chdir'ed into). One of the
4273 disadvantages of using this option is that the autofs code is
4274 relatively new and the in-place mounts have not been throughly tested.
4276 If this option is set to @samp{no}, then @i{Amd}'s autofs code will
4277 use symlinks instead of lofs-type mounts for local references. This
4278 has the advantage of using simpler (more stable) code, but at the
4279 expense of negating one of autofs's big advantages: the hiding of
4280 @i{Amd}'s internal paths. Note that symlinks are not supported in all
4281 autofs implementations, especially those derived from Solaris Autofs
4282 v1. Also, on Solaris 2.6 and newer, autofs symlinks are not cached,
4283 resulting in repeated up-call requests to @i{Amd}.
4285 @c ----------------------------------------------------------------
4286 @node browsable_dirs Parameter, map_defaults Parameter, autofs_use_lofs Parameter, Common Parameters
4287 @comment node-name, next, previous, up
4288 @subsection @t{browsable_dirs} Parameter
4289 @cindex browsable_dirs Parameter
4291 (type=string, default=@samp{no}). If @samp{yes}, then @i{Amd}'s top-level
4292 mount points will be browsable to @b{readdir}(3) calls. This means you
4293 could run for example @b{ls}(1) and see what keys are available to mount
4294 in that directory. Not all entries are made visible to @b{readdir}(3):
4295 the @samp{/defaults} entry, wildcard entries, and those with a @file{/}
4296 in them are not included. If you specify @samp{full} to this option,
4297 all but the @samp{/defaults} entry will be visible. Note that if you run
4298 a command which will attempt to @b{stat}(2) the entries, such as often
4299 done by @samp{ls -l} or @samp{ls -F}, @i{Amd} will attempt to mount
4300 @i{every} entry in that map. This is often called a ``mount storm''.
4302 Note that mount storms are mostly avoided by using autofs mounts
4303 (@samp{mount_type = autofs}).
4305 @c ----------------------------------------------------------------
4306 @node map_defaults Parameter, map_options Parameter, browsable_dirs Parameter, Common Parameters
4307 @comment node-name, next, previous, up
4308 @subsection @t{map_defaults} Parameter
4309 @cindex map_defaults Parameter
4311 (type=string, default to empty). This option sets a string to be used
4312 as the map's @code{/defaults} entry, overriding any @code{/defaults}
4313 specified in the map. This allows local users to override a given
4314 map's defaults without modifying maps globally (which is impossible in
4315 sites where the maps are managed by a different administrative group).
4317 @c ----------------------------------------------------------------
4318 @node map_options Parameter, map_type Parameter, map_defaults Parameter, Common Parameters
4319 @comment node-name, next, previous, up
4320 @subsection @t{map_options} Parameter
4321 @cindex map_options Parameter
4323 (type=string, default no options). This option is the same as
4324 specifying map options on the command line to @i{Amd}, such as
4327 @c ----------------------------------------------------------------
4328 @node map_type Parameter, mount_type Parameter, map_options Parameter, Common Parameters
4329 @comment node-name, next, previous, up
4330 @subsection @t{map_type} Parameter
4331 @cindex map_type Parameter
4333 (type=string, default search all map types). If specified, @i{Amd} will
4334 initialize the map only for the type given. This is useful to avoid the
4335 default map search type used by @i{Amd} which takes longer and can have
4336 undesired side-effects such as initializing NIS even if not used.
4343 Hesiod name service from MIT
4345 Lightweight Directory Access Protocol
4347 (New) dbm style hash files
4349 Network Information Services (version 2)
4351 Network Information Services Plus (version 3)
4353 local password files
4358 @c ----------------------------------------------------------------
4359 @node mount_type Parameter, search_path Parameter, map_type Parameter, Common Parameters
4360 @comment node-name, next, previous, up
4361 @subsection @t{mount_type} Parameter
4362 @cindex mount_type Parameter
4364 (type=string, default=@samp{nfs}). All @i{Amd} mount types default to NFS.
4365 That is, @i{Amd} is an NFS server on the map mount points, for the local
4366 host it is running on. If @samp{autofs} is specified, @i{Amd} will be
4367 an autofs server for those mount points.
4369 @c ----------------------------------------------------------------
4370 @node search_path Parameter, selectors_in_defaults Parameter, mount_type Parameter, Common Parameters
4371 @comment node-name, next, previous, up
4372 @subsection @t{search_path} Parameter
4373 @cindex search_path Parameter
4375 (type=string, default no search path). This provides a
4376 (colon-delimited) search path for file maps. Using a search path,
4377 sites can allow for local map customizations and overrides, and can
4378 distributed maps in several locations as needed.
4380 @c ----------------------------------------------------------------
4381 @node selectors_in_defaults Parameter, sun_map_syntax Parameter, search_path Parameter, Common Parameters
4382 @comment node-name, next, previous, up
4383 @subsection @t{selectors_in_defaults} Parameter
4384 @cindex selectors_in_defaults Parameter
4386 (type=boolean, default=@samp{no}). If @samp{yes}, then the
4387 @samp{/defaults} entry of maps will search for and process any
4388 selectors before setting defaults for all other keys in that map.
4389 Useful when you want to set different options for a complete map based
4390 on some parameters. For example, you may want to better the NFS
4391 performance over slow slip-based networks as follows:
4395 wire==slip-net;opts:=intr,rsize=1024,wsize=1024 \
4396 wire!=slip-net;opts:=intr,rsize=8192,wsize=8192
4399 Deprecated form: selectors_on_default.
4401 @c ----------------------------------------------------------------
4402 @node sun_map_syntax Parameter, , selectors_in_defaults Parameter, Common Parameters
4403 @comment node-name, next, previous, up
4404 @subsection @t{sun_map_syntax} Parameter
4405 @cindex sun_map_syntax Parameter
4407 (type=boolean, default=@samp{no}). If @samp{yes}, then @i{Amd} will
4408 parse the map according to the Sun Automount syntax.
4411 @c ================================================================
4412 @node Global Parameters, Regular Map Parameters, Common Parameters, Amd Configuration File
4413 @comment node-name, next, previous, up
4414 @section Global Parameters
4415 @cindex amd.conf global parameters
4417 The following parameters are applicable to the @samp{[global]} section only.
4421 * auto_attrcache Parameter::
4422 * auto_dir Parameter::
4423 * cache_duration Parameter::
4424 * cluster Parameter::
4425 * debug_mtab_file Parameter::
4426 * debug_options Parameter::
4427 * dismount_interval Parameter::
4428 * domain_strip Parameter::
4429 * exec_map_timeout Parameter::
4430 * forced_unmounts Parameter::
4431 * full_os Parameter::
4432 * fully_qualified_hosts Parameter::
4433 * hesiod_base Parameter::
4435 * ldap_base Parameter::
4436 * ldap_cache_maxmem Parameter::
4437 * ldap_cache_seconds Parameter::
4438 * ldap_hostports Parameter::
4439 * ldap_proto_version Parameter::
4440 * local_domain Parameter::
4441 * localhost_address Parameter::
4442 * log_file Parameter::
4443 * log_options Parameter::
4444 * map_reload_interval Parameter::
4445 * nfs_allow_any_interface Parameter::
4446 * nfs_allow_insecure_port Parameter::
4447 * nfs_proto Parameter::
4448 * nfs_retransmit_counter Parameter::
4449 * nfs_retransmit_counter_udp Parameter::
4450 * nfs_retransmit_counter_tcp Parameter::
4451 * nfs_retransmit_counter_toplvl Parameter::
4452 * nfs_retry_interval Parameter::
4453 * nfs_retry_interval_udp Parameter::
4454 * nfs_retry_interval_tcp Parameter::
4455 * nfs_retry_interval_toplvl Parameter::
4456 * nfs_vers Parameter::
4457 * nis_domain Parameter::
4458 * normalize_hostnames Parameter::
4459 * normalize_slashes Parameter::
4462 * pid_file Parameter::
4464 * portmap_program Parameter::
4465 * preferred_amq_port Parameter::
4466 * print_pid Parameter::
4467 * print_version Parameter::
4468 * restart_mounts Parameter::
4469 * show_statfs_entries Parameter::
4470 * truncate_log Parameter::
4471 * unmount_on_exit Parameter::
4472 * use_tcpwrappers Parameter::
4473 * vendor Parameter::
4476 @c ----------------------------------------------------------------
4477 @node arch Parameter, auto_attrcache Parameter, Global Parameters, Global Parameters
4478 @comment node-name, next, previous, up
4479 @subsection @t{arch} Parameter
4480 @cindex arch Parameter
4482 (type=string, default to compiled in value). Same as the @code{-A}
4483 option to @i{Amd}. Allows you to override the value of the @i{arch}
4486 @c ----------------------------------------------------------------
4487 @node auto_attrcache Parameter, auto_dir Parameter, arch Parameter, Global Parameters
4488 @comment node-name, next, previous, up
4489 @subsection @t{auto_attrcache} Parameter
4490 @cindex auto_attrcache Parameter
4492 (type=numeric, default=0). Specify in seconds (or units of 0.1
4493 seconds, depending on the OS), what is the (kernel-side) NFS attribute
4494 cache timeout for @i{Amd}'s own automount points. A value of 0 is
4495 supposed to turn off attribute caching, meaning that @i{Amd} will be
4496 consulted via a kernel-RPC each time someone stat()'s the mount point
4497 (which could be abused as a denial-of-service attack).
4499 @emph{WARNING}: @i{Amd} depends on being able to turn off the NFS
4500 attribute cache of the client OS. If it cannot be turned off, then
4501 users may get ESTALE errors or symlinks that point to the wrong
4502 places. This is more likely under heavy use of @i{Amd}, for example
4503 if your system is experiencing frequent map changes or frequent
4504 mounts/unmounts. Therefore, under normal circumstances, this
4505 parameter should remain set to 0, to ensure that the attribute cache
4508 Unfortunately, some kernels (e.g., certain BSDs) don't have a way to
4509 turn off the NFS attribute cache. Setting this parameter to 0 is
4510 supposed to turn off attribute caching entirely, but unfortunately it
4511 does not; instead, the attribute cache is set to some internal
4512 hard-coded default (usually anywhere from 5-30 seconds). If you
4513 suspect that your OS doesn't have a reliable way of turning off the
4514 attribute cache, then it is better to set this parameter to the
4515 smallest possible non-zero value (set @samp{auto_attrcache=1} in your
4516 @code{amd.conf}). This will not eliminate the problem, but reduce the
4517 risk window somewhat. The best solutions are (1) to use @i{Amd} in
4518 Autofs mode, if it's supported in your OS, and (2) talk to your OS
4519 vendor to support a true @samp{noac} flag. See the
4520 @uref{http://www.am-utils.org/docs/am-utils/attrcache.txt,README.attrcache}
4521 document for more details.
4523 If you are able to turn off the attribute cache on your OS, alas,
4524 @i{Amd}'s performance may degrade (when not using Autofs) because
4525 every traversal of an automounter-controlled pathname will result in a
4526 lookup request from the kernel to @i{Amd}. Under heavy loads, for
4527 example when using recursive tools like @samp{find}, @samp{rdist}, or
4528 @samp{rsync}, this performance degradation can be noticeable. There
4529 are two possible solutions that some administrators have chosen to
4530 improve performance:
4535 First, you can turn off unmounting using the @samp{nounmount} mount
4536 option. This will ensure that no @i{Amd} symlink could ever change,
4537 thereby the kernel's attribute cache and @i{Amd} will always be in
4538 sync. However, this method will cause the number of mounts to keep
4539 growing, even if some are no longer in use; this has the disadvantage
4540 that your system could be more susceptible to hangs if even one of
4541 those accumulating mounts hangs due to a downed server.
4544 Second, you can turn on attribute caching carefully by setting a small
4545 automounter attribute cache value (say, one second), and a relatively
4546 large dismount interval (say, one hour). (@xref{dismount_interval
4547 Parameter}.) For example, you can set this in your @code{amd.conf}:
4552 dismount_interval = 3600
4555 This has the benefit of using the kernel's attribute cache and thus
4556 improving performance. The disadvantage with this option is that the
4557 window of vulnerability is not eliminated entirely: it is only made
4562 @c ----------------------------------------------------------------
4563 @node auto_dir Parameter, cache_duration Parameter, auto_attrcache Parameter, Global Parameters
4564 @comment node-name, next, previous, up
4565 @subsection @t{auto_dir} Parameter
4566 @cindex auto_dir Parameter
4568 (type=string, default=@samp{/a}). Same as the @code{-a} option to @i{Amd}.
4569 This sets the private directory where @i{Amd} will create
4570 sub-directories for its real mount points.
4572 @c ----------------------------------------------------------------
4573 @node cache_duration Parameter, cluster Parameter, auto_dir Parameter, Global Parameters
4574 @comment node-name, next, previous, up
4575 @subsection @t{cache_duration} Parameter
4576 @cindex cache_duration Parameter
4578 (type=numeric, default=300). Same as the @code{-c} option to @i{Amd}.
4579 Sets the duration in seconds that looked-up or mounted map entries
4580 remain in the cache.
4582 @c ----------------------------------------------------------------
4583 @node cluster Parameter, debug_mtab_file Parameter, cache_duration Parameter, Global Parameters
4584 @comment node-name, next, previous, up
4585 @subsection @t{cluster} Parameter
4586 @cindex cluster Parameter
4588 (type=string, default no cluster). Same as the @code{-C} option to
4589 @i{Amd}. Specifies the alternate HP-UX cluster to use.
4591 @c ----------------------------------------------------------------
4592 @node debug_mtab_file Parameter, debug_options Parameter, cluster Parameter, Global Parameters
4593 @comment node-name, next, previous, up
4594 @subsection @t{debug_mtab_file} Parameter
4595 @cindex debug_mtab_file Parameter
4597 (type=string, default="/tmp/mtab"). Path to mtab file that is used
4598 by @i{Amd} to store a list of mounted file systems during debug-mtab mode.
4599 This option only applies to systems that store mtab information on disk.
4601 @c ----------------------------------------------------------------
4602 @node debug_options Parameter, dismount_interval Parameter, debug_mtab_file Parameter, Global Parameters
4603 @comment node-name, next, previous, up
4604 @subsection @t{debug_options} Parameter
4605 @cindex debug_options Parameter
4607 (type=string, default no debug options). Same as the @code{-D} option
4608 to @i{Amd}. Specify any debugging options for @i{Amd}. Works only if
4609 am-utils was configured for debugging using the @code{--enable-debug}
4610 option. The additional @samp{mem} option can be turned on via
4611 @code{--enable-debug=mem}. Otherwise debugging options are ignored.
4612 Options are comma delimited, and can be preceded by the string
4613 @samp{no} to negate their meaning. You can get the list of supported
4614 debugging and logging options by running @code{amd -H}. Possible
4615 values those listed for the -D option. @xref{-D Option}.
4617 @c ----------------------------------------------------------------
4618 @node dismount_interval Parameter, domain_strip Parameter, debug_options Parameter, Global Parameters
4619 @comment node-name, next, previous, up
4620 @subsection @t{dismount_interval} Parameter
4621 @cindex dismount_interval Parameter
4623 (type=numeric, default=120). Same as the @code{-w} option to
4624 @i{Amd}. Specify in seconds, the time between attempts to dismount file
4625 systems that have exceeded their cached times.
4627 @c ----------------------------------------------------------------
4628 @node domain_strip Parameter, exec_map_timeout Parameter, dismount_interval Parameter, Global Parameters
4629 @comment node-name, next, previous, up
4630 @subsection @t{domain_strip} Parameter
4631 @cindex domain_strip Parameter
4633 (type=boolean, default=@samp{yes}). If @samp{yes}, then the domain
4634 name part referred to by @code{$@{rhost@}} is stripped off. This is
4635 useful to keep logs and smaller. If @samp{no}, then the domain name
4636 part is left changed. This is useful when using multiple domains with
4637 the same maps (as you may have hosts whose domain-stripped name is
4640 @c ----------------------------------------------------------------
4641 @node exec_map_timeout Parameter, forced_unmounts Parameter, domain_strip Parameter, Global Parameters
4642 @comment node-name, next, previous, up
4643 @subsection @t{exec_map_timeout} Parameter
4644 @cindex exec_map_timeout Parameter
4646 (type=numeric, default=10). The timeout in seconds that @i{Amd} will
4647 wait for an executable map program before an answer is returned from
4648 that program (or script). This value should be set to as small as
4649 possible while still allowing normal replies to be returned before the
4650 timer expires, because during the time that the executable map program
4651 is queried, @i{Amd} is essentially waiting and is thus not responding
4652 to any other queries. @xref{Executable maps}.
4654 @c ----------------------------------------------------------------
4655 @node forced_unmounts Parameter, full_os Parameter, exec_map_timeout Parameter, Global Parameters
4656 @comment node-name, next, previous, up
4657 @subsection @t{forced_unmounts} Parameter
4658 @cindex forced_unmounts Parameter
4660 (type=boolean, default=@samp{no}).
4661 Sometimes, mount points are hung due to unrecoverable conditions, such
4662 as when NFS servers migrate, change their IP address, are down
4663 permanently, or due to hardware failures, and more. In this case,
4664 attempting to unmount an existing mount point, or even just to
4665 @b{stat}(2) it, results in one of three fatal errors: EIO, ESTALE, or
4666 EBUSY. At that point, @i{Amd} can do little to recover that hung
4667 point (in fact, the OS cannot automatically recover either). For that
4668 reason, some OSs support special kinds of forced unmounts, which must
4669 be used very carefully: they will force an unmount immediately (or
4670 lazily on Linux), which could result in application data loss.
4671 However, that may be the only way to recover the entire host (without
4672 rebooting). Once a hung mount point is forced out, @i{Amd} can then
4673 re-mount a replacement one (if available), bringing a mostly-hung
4674 system back to operation and avoiding a potentially costly reboot.
4676 If the @samp{forced_unmounts} option is set to @samp{yes}, and the
4677 client OS supports forced or lazy unmounts, then @i{Amd} will attempt
4678 to use them if it gets any of the three serious error conditions
4679 listed above. Note that @i{Amd} will force the unmount of mount
4680 points that returned EBUSY only for @samp{type:=toplvl} mounts
4681 (@pxref{Top-level Filesystem}): that is, @i{Amd}'s own mount points.
4682 This is useful to recover from a previously hung @i{Amd}, and to
4683 ensure that an existing @i{Amd} can shutdown cleanly even if some
4684 processes are keeping its mount points busy (i.e., when a user's shell
4685 process uses @code{cd} to set its CWD to @i{Amd}'s own mount point).
4687 If this option is set to @samp{no} (the default), then @i{Amd} will
4688 not attempt this special recovery procedure.
4690 @c ----------------------------------------------------------------
4691 @node full_os Parameter, fully_qualified_hosts Parameter, forced_unmounts Parameter, Global Parameters
4692 @comment node-name, next, previous, up
4693 @subsection @t{full_os} Parameter
4694 @cindex full_os Parameter
4696 (type=string, default to compiled in value). The full name of the
4697 operating system, along with its version. Allows you to override the
4698 compiled-in full name and version of the operating system. Useful when
4699 the compiled-in name is not desired. For example, the full operating
4700 system name on linux comes up as @samp{linux}, but you can override it
4701 to @samp{linux-2.2.5}.
4703 @c ----------------------------------------------------------------
4704 @node fully_qualified_hosts Parameter, hesiod_base Parameter, full_os Parameter, Global Parameters
4705 @comment node-name, next, previous, up
4706 @subsection @t{fully_qualified_hosts} Parameter
4707 @cindex fully_qualified_hosts Parameter
4709 (type=string, default=@samp{no}). If @samp{yes}, @i{Amd} will perform RPC
4710 authentication using fully-qualified host names. This is necessary for
4711 some systems, and especially when performing cross-domain mounting. For
4712 this function to work, the @i{Amd} variable @samp{$@{hostd@}} is used,
4713 requiring that @samp{$@{domain@}} not be null.
4715 @c ----------------------------------------------------------------
4716 @node hesiod_base Parameter, karch Parameter, fully_qualified_hosts Parameter, Global Parameters
4717 @comment node-name, next, previous, up
4718 @subsection @t{hesiod_base} Parameter
4719 @cindex hesiod_base Parameter
4721 (type=string, default=@samp{automount}). Specify the base name for
4724 @c ----------------------------------------------------------------
4725 @node karch Parameter, ldap_base Parameter, hesiod_base Parameter, Global Parameters
4726 @comment node-name, next, previous, up
4727 @subsection @t{karch} Parameter
4728 @cindex karch Parameter
4730 (type=string, default to karch of the system). Same as the @code{-k}
4731 option to @i{Amd}. Allows you to override the kernel-architecture of
4732 your system. Useful for example on Sun (Sparc) machines, where you can
4733 build one @i{Amd} binary, and run it on multiple machines, yet you want
4734 each one to get the correct @i{karch} variable set (for example, sun4c,
4735 sun4m, sun4u, etc.) Note that if not specified, @i{Amd} will use
4736 @b{uname}(2) to figure out the kernel architecture of the machine.
4738 @c ----------------------------------------------------------------
4739 @node ldap_base Parameter, ldap_cache_maxmem Parameter, karch Parameter, Global Parameters
4740 @comment node-name, next, previous, up
4741 @subsection @t{ldap_base} Parameter
4742 @cindex ldap_base Parameter
4744 (type=string, default not set).
4745 Specify the base name for LDAP. This often includes LDAP-specific
4746 values such as country and organization.
4748 @c ----------------------------------------------------------------
4749 @node ldap_cache_maxmem Parameter, ldap_cache_seconds Parameter, ldap_base Parameter, Global Parameters
4750 @comment node-name, next, previous, up
4751 @subsection @t{ldap_cache_maxmem} Parameter
4752 @cindex ldap_cache_maxmem Parameter
4754 (type=numeric, default=131072). Specify the maximum memory @i{Amd}
4755 should use to cache LDAP entries.
4757 @c ----------------------------------------------------------------
4758 @node ldap_cache_seconds Parameter, ldap_hostports Parameter, ldap_cache_maxmem Parameter, Global Parameters
4759 @comment node-name, next, previous, up
4760 @subsection @t{ldap_cache_seconds} Parameter
4761 @cindex ldap_cache_seconds Parameter
4763 (type=numeric, default=0). Specify the number of seconds to keep
4764 entries in the cache.
4766 @c ----------------------------------------------------------------
4767 @node ldap_hostports Parameter, ldap_proto_version Parameter, ldap_cache_seconds Parameter, Global Parameters
4768 @comment node-name, next, previous, up
4769 @subsection @t{ldap_hostports} Parameter
4770 @cindex ldap_hostports Parameter
4772 (type=string, default not set).
4773 Specify the LDAP host and port values.
4775 @c ----------------------------------------------------------------
4776 @node ldap_proto_version Parameter, local_domain Parameter, ldap_hostports Parameter, Global Parameters
4777 @comment node-name, next, previous, up
4778 @subsection @t{ldap_proto_version} Parameter
4779 @cindex ldap_proto_version Parameter
4781 (type=numeric, default=2). Specify the LDAP protocol version to use.
4782 With a value of 3 will use LDAPv3 protocol.
4784 @c ----------------------------------------------------------------
4785 @node local_domain Parameter, localhost_address Parameter, ldap_proto_version Parameter, Global Parameters
4786 @comment node-name, next, previous, up
4787 @subsection @t{local_domain} Parameter
4788 @cindex local_domain Parameter
4790 (type=string, default no sub-domain). Same as the @code{-d} option
4791 to @i{Amd}. Specify the local domain name. If this option is not given
4792 the domain name is determined from the hostname, by removing the first
4793 component of the fully-qualified host name.
4795 @c ----------------------------------------------------------------
4796 @node localhost_address Parameter, log_file Parameter, local_domain Parameter, Global Parameters
4797 @comment node-name, next, previous, up
4798 @subsection @t{localhost_address} Parameter
4799 @cindex localhost_address Parameter
4801 (type=string, default to localhost or 127.0.0.1). Specify the name or
4802 IP address for @i{Amd} to use when connecting the sockets for the
4803 local NFS server and the RPC server. This defaults to 127.0.0.1 or
4804 whatever the host reports as its local address. This parameter is
4805 useful on hosts with multiple addresses where you want to force
4806 @i{Amd} to connect to a specific address.
4808 @c ----------------------------------------------------------------
4809 @node log_file Parameter, log_options Parameter, localhost_address Parameter, Global Parameters
4810 @comment node-name, next, previous, up
4811 @subsection @t{log_file} Parameter
4812 @cindex log_file Parameter
4814 (type=string, default=@samp{stderr}). Same as the @code{-l} option to
4815 @i{Amd}. Specify a file name to log @i{Amd} events to.
4816 If the string @samp{/dev/stderr} is specified,
4817 @i{Amd} will send its events to the standard error file descriptor.
4819 If the string @samp{syslog} is given, @i{Amd} will record its events
4820 with the system logger @b{syslogd}(8). If your system supports syslog
4821 facilities, then the default facility used is @samp{LOG_DAEMON}.
4823 When using syslog, if you wish to change the facility, append its name
4824 to the option name, delimited by a single colon. For example, if it is
4825 the string @samp{syslog:local7} then @i{Amd} will log messages via
4826 @b{syslog}(3) using the @samp{LOG_LOCAL7} facility. If the facility
4827 name specified is not recognized, @i{Amd} will default to @samp{LOG_DAEMON}.
4828 Note: while you can use any syslog facility available on your system, it
4829 is generally a bad idea to use those reserved for other services such as
4830 @samp{kern}, @samp{lpr}, @samp{cron}, etc.
4832 @c ----------------------------------------------------------------
4833 @node log_options Parameter, map_reload_interval Parameter, log_file Parameter, Global Parameters
4834 @comment node-name, next, previous, up
4835 @subsection @t{log_options} Parameter
4836 @cindex log_options Parameter
4838 (type=string, default=``defaults''). Same as the @code{-x}
4839 option to @i{Amd}. Specify any logging options for @i{Amd}. Options
4840 are comma delimited, and can be preceded by the string @samp{no} to
4841 negate their meaning. The @samp{debug} logging option is only available
4842 if am-utils was configured with @code{--enable-debug}. You can get the
4843 list of supported debugging options by running @code{amd -H}. Possible
4850 an alias for "fatal,error,user,warning,info"
4854 non-fatal system errors (cannot be turned off)
4856 fatal errors (cannot be turned off)
4862 additional statistical information
4864 non-fatal user errors
4871 @c ----------------------------------------------------------------
4872 @node map_reload_interval Parameter, nfs_allow_any_interface Parameter, log_options Parameter, Global Parameters
4873 @comment node-name, next, previous, up
4874 @subsection @t{map_reload_interval} Parameter
4875 @cindex map_reload_interval Parameter
4877 (type=numeric, default=3600). The number of seconds that @i{Amd} will
4878 wait before it checks to see if any maps have changed at their source
4879 (NIS servers, LDAP servers, files, etc.). @i{Amd} will reload only
4880 those maps that have changed.
4882 @c ----------------------------------------------------------------
4883 @node nfs_allow_any_interface Parameter, nfs_allow_insecure_port Parameter, map_reload_interval Parameter, Global Parameters
4884 @comment node-name, next, previous, up
4885 @subsection @t{nfs_allow_any_interface} Parameter
4886 @cindex nfs_allow_any_interface Parameter
4888 (type=string, default=@samp{no}). Normally @i{Amd} accepts local NFS
4889 packets only from 127.0.0.1. If this parameter is set to @samp{yes},
4890 then @i{amd} will accept local NFS packets from any local interface;
4891 this is useful on hosts that may have multiple interfaces where the
4892 system is forced to send all outgoing packets (even those bound to the
4893 same host) via an address other than 127.0.0.1.
4895 @c ----------------------------------------------------------------
4896 @node nfs_allow_insecure_port Parameter, nfs_proto Parameter, nfs_allow_any_interface Parameter, Global Parameters
4897 @comment node-name, next, previous, up
4898 @subsection @t{nfs_allow_insecure_port} Parameter
4899 @cindex nfs_allow_insecure_port Parameter
4901 (type=string, default=@samp{no}). Normally @i{Amd} will refuse requests
4902 coming from unprivileged ports (i.e., ports >= 1024 on Unix systems),
4903 so that only privileged users and the kernel can send NFS requests to
4904 it. However, some kernels (certain versions of Darwin, MacOS X, and
4905 Linux) have bugs that cause them to use unprivileged ports in certain
4906 situations, which causes @i{Amd} to stop dead in its tracks. This
4907 parameter allows @i{Amd} to operate normally even on such systems, at the
4908 expense of a slight decrease in the security of its operations. If
4909 you see messages like ``ignoring request from foo:1234, port not
4910 reserved'' in your @i{Amd} log, try enabling this parameter and give it
4913 @c ----------------------------------------------------------------
4914 @node nfs_proto Parameter, nfs_retransmit_counter Parameter, nfs_allow_insecure_port Parameter, Global Parameters
4915 @comment node-name, next, previous, up
4916 @subsection @t{nfs_proto} Parameter
4917 @cindex nfs_proto Parameter
4919 (type=string, default to trying version tcp then udp). By default,
4920 @i{Amd} tries @code{tcp} and then @code{udp}. This option forces the
4921 overall NFS protocol used to TCP or UDP. It overrides what is in the
4922 @i{Amd} maps, and is useful when @i{Amd} is compiled with TCP support
4923 in NFSv2/NFSv3 that may not be stable. With this option you can turn
4924 off the complete usage of TCP for NFS dynamically (without having to
4925 recompile @i{Amd}), and use UDP only, until such time as TCP support
4928 @c ----------------------------------------------------------------
4929 @node nfs_retransmit_counter Parameter, nfs_retransmit_counter_udp Parameter, nfs_proto Parameter, Global Parameters
4930 @comment node-name, next, previous, up
4931 @subsection @t{nfs_retransmit_counter} Parameter
4932 @cindex nfs_retransmit_counter Parameter
4934 (type=numeric, default=11). Same as the @i{retransmit} part of the
4935 @code{-t} @i{timeout.retransmit} option to @i{Amd}. Specifies the
4936 number of NFS retransmissions that the kernel will use to communicate
4937 with @i{Amd} using either UDP or TCP mounts. @xref{-t Option}.
4939 @c ----------------------------------------------------------------
4940 @node nfs_retransmit_counter_udp Parameter, nfs_retransmit_counter_tcp Parameter, nfs_retransmit_counter Parameter, Global Parameters
4941 @comment node-name, next, previous, up
4942 @subsection @t{nfs_retransmit_counter_udp} Parameter
4943 @cindex nfs_retransmit_counter_udp Parameter
4944 @cindex nfs_retransmit_counter Parameter
4947 (type=numeric, default=11). Same as the @i{nfs_retransmit_counter}
4948 parameter, but applied globally only to UDP mounts.
4949 @xref{nfs_retransmit_counter Parameter}.
4951 @c ----------------------------------------------------------------
4952 @node nfs_retransmit_counter_tcp Parameter, nfs_retransmit_counter_toplvl Parameter, nfs_retransmit_counter_udp Parameter, Global Parameters
4953 @comment node-name, next, previous, up
4954 @subsection @t{nfs_retransmit_counter_tcp} Parameter
4955 @cindex nfs_retransmit_counter_tcp Parameter
4956 @cindex nfs_retransmit_counter Parameter
4959 (type=numeric, default=11). Same as the @i{nfs_retransmit_counter}
4960 parameter, but applied globally only to TCP mounts.
4961 @xref{nfs_retransmit_counter Parameter}.
4963 @c ----------------------------------------------------------------
4964 @node nfs_retransmit_counter_toplvl Parameter, nfs_retry_interval Parameter, nfs_retransmit_counter_tcp Parameter, Global Parameters
4965 @comment node-name, next, previous, up
4966 @subsection @t{nfs_retransmit_counter_toplvl} Parameter
4967 @cindex nfs_retransmit_counter_toplvl Parameter
4968 @cindex nfs_retransmit_counter Parameter
4971 (type=numeric, default=11). Same as the @i{nfs_retransmit_counter}
4972 parameter, applied only for @i{Amd}'s top-level UDP mounts. On some
4973 systems it is useful to set this differently than the OS default, so
4974 as to better tune @i{Amd}'s responsiveness under heavy scheduler
4975 loads. @xref{nfs_retransmit_counter Parameter}.
4977 @c ----------------------------------------------------------------
4978 @node nfs_retry_interval Parameter, nfs_retry_interval_udp Parameter, nfs_retransmit_counter_toplvl Parameter, Global Parameters
4979 @comment node-name, next, previous, up
4980 @subsection @t{nfs_retry_interval} Parameter
4981 @cindex nfs_retry_interval Parameter
4983 (type=numeric, default=8). Same as the @i{timeout} part of the
4984 @code{-t} @i{timeout.retransmit} option to @i{Amd}. Specifies the NFS
4985 timeout interval, in @emph{tenths} of seconds, between NFS/RPC retries
4986 (for UDP or TCP). This is the value that the kernel will use to
4987 communicate with @i{Amd}. @xref{-t Option}.
4989 @i{Amd} relies on the kernel RPC retransmit mechanism to trigger mount
4990 retries. The values of the @i{nfs_retransmit_counter} and the
4991 @i{nfs_retry_interval} parameters change the overall retry interval.
4992 Too long an interval gives poor interactive response; too short an
4993 interval causes excessive retries.
4995 @c ----------------------------------------------------------------
4996 @node nfs_retry_interval_udp Parameter, nfs_retry_interval_tcp Parameter, nfs_retry_interval Parameter, Global Parameters
4997 @comment node-name, next, previous, up
4998 @subsection @t{nfs_retry_interval_udp} Parameter
4999 @cindex nfs_retry_interval_udp Parameter
5000 @cindex nfs_retry_interval Parameter
5003 (type=numeric, default=8). Same as the @i{nfs_retry_interval}
5004 parameter, but applied globally only to UDP mounts.
5005 @xref{nfs_retry_interval Parameter}.
5007 @c ----------------------------------------------------------------
5008 @node nfs_retry_interval_tcp Parameter, nfs_retry_interval_toplvl Parameter, nfs_retry_interval_udp Parameter, Global Parameters
5009 @comment node-name, next, previous, up
5010 @subsection @t{nfs_retry_interval_tcp} Parameter
5011 @cindex nfs_retry_interval_tcp Parameter
5012 @cindex nfs_retry_interval Parameter
5015 (type=numeric, default=8). Same as the @i{nfs_retry_interval}
5016 parameter, but applied globally only to TCP mounts.
5017 @xref{nfs_retry_interval Parameter}.
5019 @c ----------------------------------------------------------------
5020 @node nfs_retry_interval_toplvl Parameter, nfs_vers Parameter, nfs_retry_interval_tcp Parameter, Global Parameters
5021 @comment node-name, next, previous, up
5022 @subsection @t{nfs_retry_interval_toplvl} Parameter
5023 @cindex nfs_retry_interval_toplvl Parameter
5024 @cindex nfs_retry_interval Parameter
5027 (type=numeric, default=8). Same as the @i{nfs_retry_interval}
5028 parameter, applied only for @i{Amd}'s top-level UDP mounts. On some
5029 systems it is useful to set this differently than the OS default, so
5030 as to better tune @i{Amd}'s responsiveness under heavy scheduler
5031 loads. @xref{nfs_retry_interval Parameter}.
5033 @c ----------------------------------------------------------------
5034 @node nfs_vers Parameter, nis_domain Parameter, nfs_retry_interval_toplvl Parameter, Global Parameters
5035 @comment node-name, next, previous, up
5036 @subsection @t{nfs_vers} Parameter
5037 @cindex nfs_vers Parameter
5039 (type=numeric, default to trying version 3 then 2). By default,
5040 @i{Amd} tries version 3 and then version 2. This option forces the
5041 overall NFS protocol used to version 3 or 2. It overrides what is in
5042 the @i{Amd} maps, and is useful when @i{Amd} is compiled with NFSv3
5043 support that may not be stable. With this option you can turn off the
5044 complete usage of NFSv3 dynamically (without having to recompile
5045 @i{Amd}), and use NFSv2 only, until such time as NFSv3 support is
5048 @c ----------------------------------------------------------------
5049 @node nis_domain Parameter, normalize_hostnames Parameter, nfs_vers Parameter, Global Parameters
5050 @comment node-name, next, previous, up
5051 @subsection @t{nis_domain} Parameter
5052 @cindex nis_domain Parameter
5054 (type=string, default to local NIS domain name). Same as the
5055 @code{-y} option to @i{Amd}. Specify an alternative NIS domain from
5056 which to fetch the NIS maps. The default is the system domain name.
5057 This option is ignored if NIS support is not available.
5059 @c ----------------------------------------------------------------
5060 @node normalize_hostnames Parameter, normalize_slashes Parameter, nis_domain Parameter, Global Parameters
5061 @comment node-name, next, previous, up
5062 @subsection @t{normalize_hostnames} Parameter
5063 @cindex normalize_hostnames Parameter
5065 (type=boolean, default=@samp{no}). Same as the @code{-n} option to @i{Amd}.
5066 If @samp{yes}, then the name referred to by @code{$@{rhost@}} is normalized
5067 relative to the host database before being used. The effect is to
5068 translate aliases into ``official'' names.
5070 @c ----------------------------------------------------------------
5071 @node normalize_slashes Parameter, os Parameter, normalize_hostnames Parameter, Global Parameters
5072 @comment node-name, next, previous, up
5073 @subsection @t{normalize_slashes} Parameter
5074 @cindex normalize_slashes Parameter
5076 (type=boolean, default=@samp{yes}). If @samp{yes} then amd will
5077 condense all multiple @code{/} (slash) characters into one and remove
5078 all trailing slashes. If @samp{no}, then amd will not touch strings
5079 that may contain repeated or trailing slashes. The latter is
5080 sometimes useful with SMB mounts, which often require multiple slash
5081 characters in pathnames.
5083 @c ----------------------------------------------------------------
5084 @node os Parameter, osver Parameter, normalize_slashes Parameter, Global Parameters
5085 @comment node-name, next, previous, up
5086 @subsection @t{os} Parameter
5087 @cindex os Parameter
5089 (type=string, default to compiled in value). Same as the @code{-O}
5090 option to @i{Amd}. Allows you to override the compiled-in name of the
5091 operating system. Useful when the built-in name is not desired for
5092 backward compatibility reasons. For example, if the built-in name is
5093 @samp{sunos5}, you can override it to @samp{sos5}, and use older maps
5094 which were written with the latter in mind.
5097 @c ----------------------------------------------------------------
5098 @node osver Parameter, pid_file Parameter, os Parameter, Global Parameters
5099 @comment node-name, next, previous, up
5100 @subsection @t{osver} Parameter
5101 @cindex osver Parameter
5103 (type=string, default to compiled in value). Same as the @code{-o}
5104 option to @i{Amd}. Allows you to override the compiled-in version
5105 number of the operating system. Useful when the built-in version is not
5106 desired for backward compatibility reasons. For example, if the build
5107 in version is @samp{2.5.1}, you can override it to @samp{5.5.1}, and use
5108 older maps that were written with the latter in mind.
5110 @c ----------------------------------------------------------------
5111 @node pid_file Parameter, plock Parameter, osver Parameter, Global Parameters
5112 @comment node-name, next, previous, up
5113 @subsection @t{pid_file} Parameter
5114 @cindex pid_file Parameter
5116 (type=string, default=@samp{/dev/stdout}). Specify a file to store the process
5117 ID of the running daemon into. If not specified, @i{Amd} will print its
5118 process id onto the standard output. Useful for killing @i{Amd} after
5119 it had run. Note that the PID of a running @i{Amd} can also be
5120 retrieved via @i{Amq} (@pxref{Amq -p option}).
5122 This file is used only if the @samp{print_pid} option is on
5123 (@pxref{print_pid Parameter}).
5125 @c ----------------------------------------------------------------
5126 @node plock Parameter, portmap_program Parameter, pid_file Parameter, Global Parameters
5127 @comment node-name, next, previous, up
5128 @subsection @t{plock} Parameter
5129 @cindex plock Parameter
5131 (type=boolean, default=@samp{yes}). Same as the @code{-S} option to @i{Amd}.
5132 If @samp{yes}, lock the running executable pages of @i{Amd} into memory.
5133 To improve @i{Amd}'s performance, systems that support the @b{plock}(3)
5135 call can lock the @i{Amd} process into memory. This way there is less
5136 chance the operating system will schedule, page out, and swap the
5137 @i{Amd} process as needed. This improves @i{Amd}'s performance, at the
5138 cost of reserving the memory used by the @i{Amd} process (making it
5139 unavailable for other processes).
5141 @c ----------------------------------------------------------------
5142 @node portmap_program Parameter, preferred_amq_port Parameter, plock Parameter, Global Parameters
5143 @comment node-name, next, previous, up
5144 @subsection @t{portmap_program} Parameter
5145 @cindex portmap_program Parameter
5147 (type=numeric, default=300019). Specify an alternate Port-mapper RPC
5148 program number, other than the official number. This is useful when
5149 running multiple @i{Amd} processes. For example, you can run another
5150 @i{Amd} in ``test'' mode, without affecting the primary @i{Amd} process
5151 in any way. For safety reasons, the alternate program numbers that can
5152 be specified must be in the range 300019-300029, inclusive. @i{Amq} has
5153 an option @code{-P} which can be used to specify an alternate program
5154 number of an @i{Amd} to contact. In this way, amq can fully control any
5155 number of @i{Amd} processes running on the same host.
5157 @c ----------------------------------------------------------------
5158 @node preferred_amq_port Parameter, print_pid Parameter, portmap_program Parameter, Global Parameters
5159 @comment node-name, next, previous, up
5160 @subsection @t{preferred_amq_port} Parameter
5161 @cindex preferred_amq_port Parameter
5163 (type=numeric, default=0). Specify an alternate Port-mapper RPC port
5164 number for @i{Amd}'s @i{Amq} service. This is used for both UDP and
5165 TCP. Setting this value to 0 (or not defining it) will cause @i{Amd}
5166 to select an arbitrary port number. Setting the @i{Amq} RPC service
5167 port to a specific number is useful in firewalled or NAT'ed
5168 environments, where you need to know which port @i{Amd} will listen
5171 @c ----------------------------------------------------------------
5172 @node print_pid Parameter, print_version Parameter, preferred_amq_port Parameter, Global Parameters
5173 @comment node-name, next, previous, up
5174 @subsection @t{print_pid} Parameter
5175 @cindex print_pid Parameter
5177 (type=boolean, default=@samp{no}). Same as the @code{-p} option to @i{Amd}.
5178 If @samp{yes}, @i{Amd} will print its process ID upon starting.
5180 @c ----------------------------------------------------------------
5181 @node print_version Parameter, restart_mounts Parameter, print_pid Parameter, Global Parameters
5182 @comment node-name, next, previous, up
5183 @subsection @t{print_version} Parameter
5184 @cindex print_version Parameter
5186 (type=boolean, default=@samp{no}). Same as the @code{-v} option to @i{Amd},
5187 but the version prints and @i{Amd} continues to run. If @samp{yes}, @i{Amd}
5188 will print its version information string, which includes some
5189 configuration and compilation values.
5191 @c ----------------------------------------------------------------
5192 @node restart_mounts Parameter, show_statfs_entries Parameter, print_version Parameter, Global Parameters
5193 @comment node-name, next, previous, up
5194 @subsection @t{restart_mounts} Parameter
5195 @cindex restart_mounts Parameter
5197 (type=boolean, default=@samp{no}). Same as the @code{-r} option to @i{Amd}.
5198 If @samp{yes} @i{Amd} will scan the mount table to determine which file
5199 systems are currently mounted. Whenever one of these would have been
5200 auto-mounted, @i{Amd} inherits it.
5202 @c ----------------------------------------------------------------
5203 @node show_statfs_entries Parameter, truncate_log Parameter, restart_mounts Parameter, Global Parameters
5204 @comment node-name, next, previous, up
5205 @subsection @t{show_statfs_entries} Parameter
5206 @cindex show_statfs_entries Parameter
5208 (type=boolean), default=@samp{no}). If @samp{yes}, then all maps which are
5209 browsable will also show the number of entries (keys) they have when
5210 @b{df}(1) runs. (This is accomplished by returning non-zero values to
5211 the @b{statfs}(2) system call).
5213 @c ----------------------------------------------------------------
5214 @node truncate_log Parameter, unmount_on_exit Parameter, show_statfs_entries Parameter, Global Parameters
5215 @comment node-name, next, previous, up
5216 @subsection @t{truncate_log} Parameter
5217 @cindex truncate_log Parameter
5219 (type=boolean), default=@samp{no}). If @samp{yes}, then @i{Amd} will
5220 truncate the log file (if it's a regular file) on startup. This could
5221 be useful when conducting extensive testing on @i{Amd} maps (or
5222 @i{Amd} itself) and you don't want to see log data from a previous run
5225 @c ----------------------------------------------------------------
5226 @node unmount_on_exit Parameter, use_tcpwrappers Parameter, truncate_log Parameter, Global Parameters
5227 @comment node-name, next, previous, up
5228 @subsection @t{unmount_on_exit} Parameter
5229 @cindex unmount_on_exit Parameter
5231 (type=boolean, default=@samp{no}). If @samp{yes}, then @i{Amd} will attempt
5232 to unmount all file systems which it knows about. Normally it leaves
5233 all (esp. NFS) mounted file systems intact. Note that @i{Amd} does not
5234 know about file systems mounted before it starts up, unless the
5235 @samp{restart_mounts} option is used (@pxref{restart_mounts Parameter}).
5237 @c ----------------------------------------------------------------
5238 @node use_tcpwrappers Parameter, vendor Parameter, unmount_on_exit Parameter, Global Parameters
5239 @comment node-name, next, previous, up
5240 @subsection @t{use_tcpwrappers} Parameter
5241 @cindex use_tcpwrappers Parameter
5243 (type=boolean), default=@samp{yes}). If @samp{yes}, then amd will use
5244 the tcpwrappers (tcpd/librwap) library (if available) to control
5245 access to @i{Amd} via the @code{/etc/hosts.allow} and
5246 @code{/etc/hosts.deny} files. @i{Amd} will verify that the host
5247 running @i{Amq} is authorized to connect. The @code{amd} service name
5248 must used in the @code{/etc/hosts.allow} and @code{/etc/hosts.deny}
5249 files. For example, to allow only localhost to connect to @i{Amd},
5250 add this line to @code{/etc/hosts.allow}:
5256 and this line to @code{/etc/hosts.deny}:
5262 Consult the man pages for @b{hosts_access}(5) for more information on using
5263 the tcpwrappers access-control library.
5265 Note that in particular, you should not configure your @code{hosts.allow}
5266 file to spawn a command for @i{Amd}: that will cause @i{Amd} to not be able
5267 to @code{waitpid} on the child process ID of any background un/mount that
5268 @i{Amd} issued, resulting in a confused @i{Amd} that does not know what
5269 happened to those background un/mount requests.
5271 @c ----------------------------------------------------------------
5272 @node vendor Parameter, , use_tcpwrappers Parameter, Global Parameters
5273 @comment node-name, next, previous, up
5274 @subsection @t{vendor} Parameter
5275 @cindex vendor Parameter
5277 (type=string, default to compiled in value). The name of the vendor of
5278 the operating system. Overrides the compiled-in vendor name. Useful
5279 when the compiled-in name is not desired. For example, most Intel based
5280 systems set the vendor name to @samp{unknown}, but you can set it to
5283 @c ================================================================
5284 @node Regular Map Parameters, amd.conf Examples, Global Parameters, Amd Configuration File
5285 @comment node-name, next, previous, up
5286 @section Regular Map Parameters
5287 @cindex amd.conf regular map parameters
5289 The following parameters are applicable only to regular map sections.
5292 * map_name Parameter::
5296 @c ----------------------------------------------------------------
5297 @node map_name Parameter, tag Parameter, Regular Map Parameters, Regular Map Parameters
5298 @comment node-name, next, previous, up
5299 @subsection map_name Parameter
5300 @cindex map_name Parameter
5302 (type=string, must be specified). Name of the map where the keys are
5305 @c ----------------------------------------------------------------
5306 @node tag Parameter, , map_name Parameter, Regular Map Parameters
5307 @comment node-name, next, previous, up
5308 @subsection tag Parameter
5309 @cindex tag Parameter
5311 (type=string, default no tag). Each map entry in the configuration file
5312 can be tagged. If no tag is specified, that map section will always be
5313 processed by @i{Amd}. If it is specified, then @i{Amd} will process the map
5314 if the @code{-T} option was given to @i{Amd}, and the value given to that
5315 command-line option matches that in the map section.
5317 @c ================================================================
5318 @node amd.conf Examples, , Regular Map Parameters, Amd Configuration File
5319 @comment node-name, next, previous, up
5320 @section amd.conf Examples
5321 @cindex amd.conf examples
5323 The following is the actual @code{amd.conf} file I used at the
5324 Computer Science Department of Columbia University.
5327 # GLOBAL OPTIONS SECTION
5329 normalize_hostnames = no
5331 #pid_file = /var/run/amd.pid
5332 restart_mounts = yes
5333 #unmount_on_exit = yes
5335 log_file = /var/log/amd
5337 #debug_options = defaults
5339 selectors_in_defaults = yes
5340 # config.guess picks up "sunos5" and I don't want to edit my maps yet
5342 # if you print_version after setting up "os", it will show it.
5345 search_path = /etc/amdmaps:/usr/lib/amd:/usr/local/AMD/lib
5346 browsable_dirs = yes
5347 fully_qualified_hosts = no
5349 # DEFINE AN AMD MOUNT POINT
5363 map_name = amd.import
5367 map_name = amd.tftpboot
5370 @c ################################################################
5371 @node Run-time Administration, FSinfo, Amd Configuration File, Top
5372 @comment node-name, next, previous, up
5373 @chapter Run-time Administration
5374 @cindex Run-time administration
5384 @node Starting Amd, Stopping Amd, Run-time Administration, Run-time Administration
5385 @comment node-name, next, previous, up
5386 @section Starting @i{Amd}
5387 @cindex Starting Amd
5388 @cindex Additions to /etc/rc.local
5389 @cindex /etc/rc.local additions
5392 @i{Amd} is best started from @samp{/etc/rc.local} on BSD systems, or
5393 from the appropriate start-level script in @samp{/etc/init.d} on System V
5397 if [ -f /usr/local/sbin/ctl-amd ]; then
5398 /usr/local/sbin/ctl-amd start; (echo -n ' amd') > /dev/console
5403 The shell script, @samp{ctl-amd} is used to start, stop, or restart
5404 @i{Amd}. It is a relatively generic script. All options you want to
5405 set should not be made in this script, but rather updated in the
5406 @file{amd.conf} file. @xref{Amd Configuration File}.
5408 If you do not wish to use an @i{Amd} configuration file, you may start
5409 @i{Amd} manually. For example, getting the map entries via NIS:
5412 amd -r -l /var/log/amd `ypcat -k auto.master`
5415 @node Stopping Amd, Restarting Amd, Starting Amd, Run-time Administration
5416 @comment node-name, next, previous, up
5417 @section Stopping @i{Amd}
5418 @cindex Stopping Amd
5419 @cindex SIGTERM signal
5420 @cindex SIGINT signal
5422 @i{Amd} stops in response to two signals.
5426 causes the top-level automount points to be unmounted and then @i{Amd}
5427 to exit. Any automounted filesystems are left mounted. They can be
5428 recovered by restarting @i{Amd} with the @code{-r} command line option.@refill
5431 causes @i{Amd} to attempt to unmount any filesystems which it has
5432 automounted, in addition to the actions of @samp{SIGTERM}. This signal
5433 is primarily used for debugging.@refill
5436 Actions taken for other signals are undefined.
5438 The easiest and safest way to stop @i{Amd}, without having to find its
5439 process ID by hand, is to use the @file{ctl-amd} script, as with:
5445 @node Restarting Amd, Controlling Amd, Stopping Amd, Run-time Administration
5446 @comment node-name, next, previous, up
5447 @section Restarting @i{Amd}
5448 @cindex Restarting Amd
5449 @cindex Killing and starting Amd
5451 Before @i{Amd} can be started, it is vital to ensure that no other
5452 @i{Amd} processes are managing any of the mount points, and that the
5453 previous process(es) have terminated cleanly. When a terminating signal
5454 is set to @i{Amd}, the automounter does @emph{not} terminate right then.
5455 Rather, it starts by unmounting all of its managed mount mounts in the
5456 background, and then terminates. It usually takes a few seconds for
5457 this process to happen, but it can take an arbitrarily longer time. If
5458 two or more @i{Amd} processes attempt to manage the same mount point, it
5459 usually will result in a system lockup.
5461 The easiest and safest way to restart @i{Amd}, without having to find
5462 its process ID by hand, sending it the @samp{SIGTERM} signal, waiting for @i{Amd}
5463 to die cleanly, and verifying so, is to use the @file{ctl-amd} script,
5470 The script will locate the process ID of @i{Amd}, kill it, and wait for
5471 it to die cleanly before starting a new instance of the automounter.
5472 @file{ctl-amd} will wait for a total of 30 seconds for @i{Amd} to die,
5473 and will check once every 5 seconds if it had.
5475 @node Controlling Amd, , Restarting Amd, Run-time Administration
5476 @comment node-name, next, previous, up
5477 @section Controlling @i{Amd}
5478 @cindex Controlling Amd
5479 @cindex Discovering what is going on at run-time
5480 @cindex Listing currently mounted filesystems
5482 It is sometimes desirable or necessary to exercise external control
5483 over some of @i{Amd}'s internal state. To support this requirement,
5484 @i{Amd} implements an RPC interface which is used by the @dfn{Amq} program.
5485 A variety of information is available.
5487 @i{Amq} generally applies an operation, specified by a single letter option,
5488 to a list of mount points. The default operation is to obtain statistics
5489 about each mount point. This is similar to the output shown above
5490 but includes information about the number and type of accesses to each
5494 * Amq default:: Default command behavior.
5495 * Amq -f option:: Flushing the map cache.
5496 * Amq -h option:: Controlling a non-local host.
5497 * Amq -H option:: Print help message.
5498 * Amq -l option:: Controlling the log file.
5499 * Amq -m option:: Obtaining mount statistics.
5500 * Amq -p option:: Getting Amd's process ID.
5501 * Amq -P option:: Contacting alternate Amd processes.
5502 * Amq -q option:: Suppress synchronous unmounting errors.
5503 * Amq -s option:: Obtaining global statistics.
5504 * Amq -T option:: Use TCP transport.
5505 * Amq -U option:: Use UDP transport.
5506 * Amq -u option:: Forcing volumes to time out.
5507 * Amq -v option:: Version information.
5508 * Amq -w option:: Print Amd current working directory.
5509 * Other Amq options:: Three other special options.
5512 @c ----------------------------------------------------------------
5513 @node Amq default, Amq -f option, Controlling Amd, Controlling Amd
5514 @comment node-name, next, previous, up
5515 @subsection @i{Amq} default information
5517 With no arguments, @dfn{Amq} obtains a brief list of all existing
5518 mounts created by @i{Amd}. This is different from the list displayed by
5519 @b{df}(1) since the latter only includes system mount points.
5522 The output from this option includes the following information:
5526 the automount point,
5528 the filesystem type,
5530 the mount map or mount information,
5532 the internal, or system mount point.
5539 / root "root" sky:(pid75)
5540 /homes toplvl /usr/local/etc/amd.homes /homes
5541 /home toplvl /usr/local/etc/amd.home /home
5542 /homes/jsp nfs charm:/home/charm /a/charm/home/charm/jsp
5543 /homes/phjk nfs toytown:/home/toytown /a/toytown/home/toytown/ai/phjk
5547 If an argument is given then statistics for that volume name will
5548 be output. For example:
5551 What Uid Getattr Lookup RdDir RdLnk Statfs Mounted@@
5552 /homes 0 1196 512 22 0 30 90/09/14 12:32:55
5553 /homes/jsp 0 0 0 0 1180 0 90/10/13 12:56:58
5564 the count of NFS @dfn{getattr} requests on this node. This should only be
5565 non-zero for directory nodes.
5568 the count of NFS @dfn{lookup} requests on this node. This should only be
5569 non-zero for directory nodes.
5572 the count of NFS @dfn{readdir} requests on this node. This should only
5573 be non-zero for directory nodes.
5576 the count of NFS @dfn{readlink} requests on this node. This should be
5577 zero for directory nodes.
5580 the count of NFS @dfn{statfs} requests on this node. This should only
5581 be non-zero for top-level automount points.
5584 the date and time the volume name was first referenced.
5587 @c ----------------------------------------------------------------
5588 @node Amq -f option, Amq -h option, Amq default, Controlling Amd
5589 @comment node-name, next, previous, up
5590 @subsection @i{Amq} @code{-f} option
5591 @cindex Flushing the map cache
5592 @cindex Map cache, flushing
5594 The @code{-f} option causes @i{Amd} to flush the internal mount map cache.
5595 This is useful for example in Hesiod maps since @i{Amd} will not
5596 automatically notice when they have been updated. The map cache can
5597 also be synchronized with the map source by using the @samp{sync} option
5598 (@pxref{Automount Filesystem}).@refill
5600 @c ----------------------------------------------------------------
5601 @node Amq -h option, Amq -H option, Amq -f option, Controlling Amd
5602 @comment node-name, next, previous, up
5603 @subsection @i{Amq} @code{-h} option
5604 @cindex Querying an alternate host
5606 By default the local host is used. In an HP-UX cluster the root server
5607 is used since that is the only place in the cluster where @i{Amd} will
5608 be running. To query @i{Amd} on another host the @code{-h} option should
5611 @c ----------------------------------------------------------------
5612 @node Amq -H option, Amq -l option, Amq -h option, Controlling Amd
5613 @comment node-name, next, previous, up
5614 @subsection @i{Amq} @code{-H} option
5615 @cindex Displaying brief help
5616 @cindex Help; showing from Amq
5618 Print a brief help and usage string.
5620 @c ----------------------------------------------------------------
5621 @node Amq -l option, Amq -m option, Amq -H option, Controlling Amd
5622 @comment node-name, next, previous, up
5623 @subsection @i{Amq} @code{-l} option
5624 @cindex Resetting the Amd log file
5625 @cindex Setting the Amd log file via Amq
5626 @cindex Log file, resetting
5628 Tell @i{Amd} to use @i{log_file} as the log file name. For security
5629 reasons, this @emph{must} be the same log file which @i{Amd} used when
5630 started. This option is therefore only useful to refresh @i{Amd}'s open
5631 file handle on the log file, so that it can be rotated and compressed
5632 via daily cron jobs.
5634 @c ----------------------------------------------------------------
5635 @node Amq -m option, Amq -p option, Amq -l option, Controlling Amd
5636 @comment node-name, next, previous, up
5637 @subsection @i{Amq} @code{-m} option
5639 The @code{-m} option displays similar information about mounted
5640 filesystems, rather than automount points. The output includes the
5641 following information:
5645 the mount information,
5649 the filesystem type,
5651 the number of references to this filesystem,
5653 the server hostname,
5655 the state of the file server,
5657 any error which has occurred.
5663 "root" truth:(pid602) root 1 localhost is up
5664 hesiod.home /home toplvl 1 localhost is up
5665 hesiod.vol /vol toplvl 1 localhost is up
5666 hesiod.homes /homes toplvl 1 localhost is up
5667 amy:/home/amy /a/amy/home/amy nfs 5 amy is up
5668 swan:/home/swan /a/swan/home/swan nfs 0 swan is up (Permission denied)
5669 ex:/home/ex /a/ex/home/ex nfs 0 ex is down
5672 When the reference count is zero the filesystem is not mounted but
5673 the mount point and server information is still being maintained
5676 @c ----------------------------------------------------------------
5678 @comment Retained for future consideration: from the description of the
5679 @comment amq -M option removed in amd 6.0.5.
5681 A future release of @i{Amd} will include code to allow the @b{mount}(8)
5682 command to mount automount points:
5685 mount -t amd /vol hesiod.vol
5688 This will then allow @i{Amd} to be controlled from the standard system
5689 filesystem mount list.
5693 @c ----------------------------------------------------------------
5694 @node Amq -p option, Amq -P option, Amq -m option, Controlling Amd
5695 @comment node-name, next, previous, up
5696 @subsection @i{Amq} @code{-p} option
5697 @cindex Process ID; Amd
5698 @cindex Amd's process ID
5702 Return the process ID of the remote or locally running @i{Amd}. Useful
5703 when you need to send a signal to the local @i{Amd} process, and would
5704 rather not have to search through the process table. This option is
5705 used in the @file{ctl-amd} script.
5707 @c ----------------------------------------------------------------
5708 @node Amq -P option, Amq -q option, Amq -p option, Controlling Amd
5709 @comment node-name, next, previous, up
5710 @subsection @i{Amq} @code{-P} option
5711 @cindex Multiple Amd processes
5712 @cindex Running multiple Amd
5713 @cindex Debugging a new Amd configuration
5714 @cindex RPC Program numbers; Amd
5716 Contact an alternate running @i{Amd} that had registered itself on a
5717 different RPC @var{program_number} and apply all other operations to
5718 that instance of the automounter. This is useful when you run multiple
5719 copies of @i{Amd}, and need to manage each one separately. If not
5720 specified, @i{Amq} will use the default program number for @i{Amd}, 300019.
5721 For security reasons, the only alternate program numbers @i{Amd} can use
5722 range from 300019 to 300029, inclusive.
5724 For example, to kill an alternate running @i{Amd}:
5727 kill `amq -p -P 300020`
5730 @c ----------------------------------------------------------------
5731 @node Amq -q option, Amq -s option, Amq -P option, Controlling Amd
5732 @comment node-name, next, previous, up
5733 @subsection @i{Amq} @code{-q} option
5734 @cindex Unmounting a filesystem
5736 Suppress any error messages produced when a synchronous unmount fails.
5737 See @ref{Amq -u option}.
5739 @c ----------------------------------------------------------------
5740 @node Amq -s option, Amq -T option, Amq -q option, Controlling Amd
5741 @comment node-name, next, previous, up
5742 @subsection @i{Amq} @code{-s} option
5743 @cindex Global statistics
5746 The @code{-s} option displays global statistics. If any other options are specified
5747 or any filesystems named then this option is ignored. For example:
5750 requests stale mount mount unmount
5751 deferred fhandles ok failed failed
5756 @item Deferred requests
5757 are those for which an immediate reply could not be constructed. For
5758 example, this would happen if a background mount was required.
5760 @item Stale filehandles
5761 counts the number of times the kernel passes a stale filehandle to @i{Amd}.
5762 Large numbers indicate problems.
5765 counts the number of automounts which were successful.
5768 counts the number of automounts which failed.
5770 @item Unmount failed
5771 counts the number of times a filesystem could not be unmounted. Very
5772 large numbers here indicate that the time between unmount attempts
5773 should be increased.
5776 @c ----------------------------------------------------------------
5777 @node Amq -T option, Amq -U option, Amq -s option, Controlling Amd
5778 @comment node-name, next, previous, up
5779 @subsection @i{Amq} @code{-T} option
5780 @cindex Forcing Amq to use a TCP transport
5781 @cindex TCP; using with Amq
5783 The @code{-T} option causes the @i{Amq} to contact @i{Amd} using the TCP
5784 transport only (connection oriented). Normally, @i{Amq} will use TCP
5785 first, and if that failed, will try UDP.
5787 @c ----------------------------------------------------------------
5788 @node Amq -U option, Amq -u option, Amq -T option, Controlling Amd
5789 @comment node-name, next, previous, up
5790 @subsection @i{Amq} @code{-U} option
5791 @cindex Forcing Amq to use a UDP transport
5792 @cindex UDP; using with Amq
5794 The @code{-U} option causes the @i{Amq} to contact @i{Amd} using the UDP
5795 transport only (connectionless). Normally, @i{Amq} will use TCP first,
5796 and if that failed, will try UDP.
5798 @c ----------------------------------------------------------------
5799 @node Amq -u option, Amq -v option, Amq -U option, Controlling Amd
5800 @comment node-name, next, previous, up
5801 @subsection @i{Amq} @code{-u} option
5802 @cindex Forcing filesystem to time out
5803 @cindex Unmounting a filesystem
5805 The @code{-u} option causes the time-to-live interval of the named
5806 mount points to be expired, thus causing an unmount attempt. This is
5807 the only safe way to unmount an automounted filesystem. If @code{-u}
5808 is repeated, then @i{Amd} will attempt to unmount the filesystem
5809 synchronously. This makes things like
5812 amq -uu /t/cd0d && eject cd0
5816 work as expected. Any error messages this might produce can be
5817 suppressed with the @code{-q} option. See @ref{Amq -q option}.
5819 @c The @code{-H} option informs @i{Amd} that the specified mount point
5820 @c has hung - as if its keepalive timer had expired.
5822 @c ----------------------------------------------------------------
5823 @node Amq -v option, Amq -w option, Amq -u option, Controlling Amd
5824 @comment node-name, next, previous, up
5825 @subsection @i{Amq} @code{-v} option
5826 @cindex Version information at run-time
5828 The @code{-v} option displays the version of @i{Amd} in a similar way to
5829 @i{Amd}'s @code{-v} option.
5831 @c ----------------------------------------------------------------
5832 @node Amq -w option, Other Amq options, Amq -v option, Controlling Amd
5833 @comment node-name, next, previous, up
5834 @subsection @i{Amq} @code{-w} option
5835 @cindex Getting real working directory
5837 The @code{-w} option translates a full pathname as returned by
5838 @b{getpwd}(3) into a short @i{Amd} pathname that goes through its mount
5839 points. This option requires that @i{Amd} is running.
5841 @c ----------------------------------------------------------------
5842 @node Other Amq options, , Amq -w option, Controlling Amd
5843 @comment node-name, next, previous, up
5844 @subsection Other @i{Amq} options
5845 @cindex Logging options via Amq
5846 @cindex Debugging options via Amq
5848 Two other operations are implemented. These modify the state of @i{Amd}
5849 as a whole, rather than any particular filesystem. The @code{-x} and
5850 @code{-D} options have exactly the same effect as @i{Amd}'s corresponding
5851 command line options.
5853 When @i{Amd} receives the @code{-x} flag, it disallows turning off the
5854 @samp{fatal} or @samp{error} flags. Both are on by default. They are
5855 mandatory so that @i{Amd} could report important errors, including
5856 errors relating to turning flags on/off.
5858 @c ################################################################
5859 @node FSinfo, Hlfsd, Run-time Administration, Top
5860 @comment node-name, next, previous, up
5863 @cindex Filesystem info package
5865 XXX: this chapter should be reviewed by someone knowledgeable with
5869 * FSinfo Overview:: Introduction to FSinfo.
5870 * Using FSinfo:: Basic concepts.
5871 * FSinfo Grammar:: Language syntax, semantics and examples.
5872 * FSinfo host definitions:: Defining a new host.
5873 * FSinfo host attributes:: Definable host attributes.
5874 * FSinfo filesystems:: Defining locally attached filesystems.
5875 * FSinfo static mounts:: Defining additional static mounts.
5876 * FSinfo automount definitions::
5877 * FSinfo Command Line Options::
5881 @node FSinfo Overview, Using FSinfo, FSinfo, FSinfo
5882 @comment node-name, next, previous, up
5883 @section @i{FSinfo} overview
5884 @cindex FSinfo overview
5886 @i{FSinfo} is a filesystem management tool. It has been designed to
5887 work with @i{Amd} to help system administrators keep track of the ever
5888 increasing filesystem namespace under their control.
5890 The purpose of @i{FSinfo} is to generate all the important standard
5891 filesystem data files from a single set of input data. Starting with a
5892 single data source guarantees that all the generated files are
5893 self-consistent. One of the possible output data formats is a set of
5894 @i{Amd} maps which can be used among the set of hosts described in the
5897 @i{FSinfo} implements a declarative language. This language is
5898 specifically designed for describing filesystem namespace and physical
5899 layouts. The basic declaration defines a mounted filesystem including
5900 its device name, mount point, and all the volumes and access
5901 permissions. @i{FSinfo} reads this information and builds an internal
5902 map of the entire network of hosts. Using this map, many different data
5903 formats can be produced including @file{/etc/fstab},
5904 @file{/etc/exports}, @i{Amd} mount maps and
5905 @file{/etc/bootparams}.@refill
5907 @node Using FSinfo, FSinfo Grammar, FSinfo Overview, FSinfo
5908 @comment node-name, next, previous, up
5909 @section Using @i{FSinfo}
5910 @cindex Using FSinfo
5912 The basic strategy when using @i{FSinfo} is to gather all the
5913 information about all disks on all machines into one set of
5914 declarations. For each machine being managed, the following data is
5921 List of all filesystems and, optionally, their mount points.
5923 Names of volumes stored on each filesystem.
5925 NFS export information for each volume.
5927 The list of static filesystem mounts.
5930 The following information can also be entered into the same
5931 configuration files so that all data can be kept in one place.
5935 List of network interfaces
5937 IP address of each interface
5939 Hardware address of each interface
5941 Dumpset to which each filesystem belongs
5946 To generate @i{Amd} mount maps, the automount tree must also be defined
5947 (@pxref{FSinfo automount definitions}). This will have been designed at
5948 the time the volume names were allocated. Some volume names will not be
5949 automounted, so @i{FSinfo} needs an explicit list of which volumes
5950 should be automounted.@refill
5952 Hostnames are required at several places in the @i{FSinfo} language. It
5953 is important to stick to either fully qualified names or unqualified
5954 names. Using a mixture of the two will inevitably result in confusion.
5956 Sometimes volumes need to be referenced which are not defined in the set
5957 of hosts being managed with @i{FSinfo}. The required action is to add a
5958 dummy set of definitions for the host and volume names required. Since
5959 the files generated for those particular hosts will not be used on them,
5960 the exact values used is not critical.
5962 @node FSinfo Grammar, FSinfo host definitions, Using FSinfo, FSinfo
5963 @comment node-name, next, previous, up
5964 @section @i{FSinfo} grammar
5965 @cindex FSinfo grammar
5966 @cindex Grammar, FSinfo
5968 @i{FSinfo} has a relatively simple grammar. Distinct syntactic
5969 constructs exist for each of the different types of data, though they
5970 share a common flavor. Several conventions are used in the grammar
5973 The notation, @i{list(}@t{xxx}@i{)}, indicates a list of zero or more
5974 @t{xxx}'s. The notation, @i{opt(}@t{xxx}@i{)}, indicates zero or one
5975 @t{xxx}. Items in double quotes, @i{eg} @t{"host"}, represent input
5976 tokens. Items in angle brackets, @i{eg} @var{<hostname>}, represent
5977 strings in the input. Strings need not be in double quotes, except to
5978 differentiate them from reserved words. Quoted strings may include the
5979 usual set of C ``@t{\}'' escape sequences with one exception: a
5980 backslash-newline-whitespace sequence is squashed into a single space
5981 character. To defeat this feature, put a further backslash at the start
5984 At the outermost level of the grammar, the input consists of a
5985 sequence of host and automount declarations. These declarations are
5986 all parsed before they are analyzed. This means they can appear in
5987 any order and cyclic host references are possible.
5990 fsinfo : @i{list(}fsinfo_attr@i{)} ;
5992 fsinfo_attr : host | automount ;
5996 * FSinfo host definitions::
5997 * FSinfo automount definitions::
6000 @node FSinfo host definitions, FSinfo host attributes, FSinfo Grammar, FSinfo
6001 @comment node-name, next, previous, up
6002 @section @i{FSinfo} host definitions
6003 @cindex FSinfo host definitions
6004 @cindex Defining a host, FSinfo
6006 A host declaration consists of three parts: a set of machine attribute
6007 data, a list of filesystems physically attached to the machine, and a
6008 list of additional statically mounted filesystems.
6011 host : "host" host_data @i{list(}filesystem@i{@i{)}} @i{list(}mount@i{@i{)}} ;
6014 Each host must be declared in this way exactly once. Such things as the
6015 hardware address, the architecture and operating system types and the
6016 cluster name are all specified within the @dfn{host data}.
6018 All the disks the machine has should then be described in the @dfn{list
6019 of filesystems}. When describing disks, you can specify what
6020 @dfn{volname} the disk/partition should have and all such entries are
6021 built up into a dictionary which can then be used for building the
6024 The @dfn{list of mounts} specifies all the filesystems that should be
6025 statically mounted on the machine.
6028 * FSinfo host attributes::
6029 * FSinfo filesystems::
6030 * FSinfo static mounts::
6033 @node FSinfo host attributes, FSinfo filesystems, FSinfo host definitions , FSinfo host definitions
6034 @comment node-name, next, previous, up
6035 @section @i{FSinfo} host attributes
6036 @cindex FSinfo host attributes
6037 @cindex Defining host attributes, FSinfo
6039 The host data, @dfn{host_data}, always includes the @dfn{hostname}. In
6040 addition, several other host attributes can be given.
6043 host_data : @var{<hostname>}
6044 | "@{" @i{list(}host_attrs@i{)} "@}" @var{<hostname>}
6047 host_attrs : host_attr "=" @var{<string>}
6051 host_attr : "config"
6058 The @dfn{hostname} is, typically, the fully qualified hostname of the
6064 host dylan.doc.ic.ac.uk
6069 @} dougal.doc.ic.ac.uk
6072 The options that can be given as host attributes are shown below.
6075 * FSinfo netif Option:: FSinfo host netif.
6076 * FSinfo config Option:: FSinfo host config.
6077 * FSinfo arch Option:: FSinfo host arch.
6078 * FSinfo os Option:: FSinfo host os.
6079 * FSinfo cluster Option:: FSinfo host cluster.
6082 @node FSinfo netif Option, FSinfo config Option, , FSinfo host attributes
6083 @comment node-name, next, previous, up
6084 @subsection netif Option
6086 This defines the set of network interfaces configured on the machine.
6087 The interface attributes collected by @i{FSinfo} are the IP address,
6088 subnet mask and hardware address. Multiple interfaces may be defined
6089 for hosts with several interfaces by an entry for each interface. The
6090 values given are sanity checked, but are currently unused for anything
6094 netif : "netif" @var{<string>} "@{" @i{list(}netif_attrs@i{)} "@}" ;
6096 netif_attrs : netif_attr "=" @var{<string>} ;
6098 netif_attr : "inaddr" | "netmask" | "hwaddr" ;
6105 inaddr = 129.31.81.37
6106 netmask = 0xfffffe00
6107 hwaddr = "08:00:20:01:a6:a5"
6113 @node FSinfo config Option, FSinfo arch Option, FSinfo netif Option, FSinfo host attributes
6114 @comment node-name, next, previous, up
6115 @subsection config Option
6116 @cindex FSinfo config host attribute
6117 @cindex config, FSinfo host attribute
6119 This option allows you to specify configuration variables for the
6120 startup scripts (@file{rc} scripts). A simple string should immediately
6126 config "NFS_SERVER=true"
6127 config "ZEPHYR=true"
6130 This option is currently unsupported.
6132 @node FSinfo arch Option, FSinfo os Option, FSinfo config Option, FSinfo host attributes
6133 @comment node-name, next, previous, up
6134 @subsection arch Option
6135 @cindex FSinfo arch host attribute
6136 @cindex arch, FSinfo host attribute
6138 This defines the architecture of the machine. For example:
6144 This is intended to be of use when building architecture specific
6145 mountmaps, however, the option is currently unsupported.
6147 @node FSinfo os Option, FSinfo cluster Option, FSinfo arch Option, FSinfo host attributes
6148 @comment node-name, next, previous, up
6149 @subsection os Option
6150 @cindex FSinfo os host attribute
6151 @cindex os, FSinfo host attribute
6153 This defines the operating system type of the host. For example:
6159 This information is used when creating the @file{fstab} files, for
6160 example in choosing which format to use for the @file{fstab} entries
6163 @node FSinfo cluster Option, , FSinfo os Option, FSinfo host attributes
6164 @comment node-name, next, previous, up
6165 @subsection cluster Option
6166 @cindex FSinfo cluster host attribute
6167 @cindex cluster, FSinfo host attribute
6169 This is used for specifying in which cluster the machine belongs. For
6176 The cluster is intended to be used when generating the automount maps,
6177 although it is currently unsupported.
6179 @node FSinfo filesystems, FSinfo static mounts, FSinfo host attributes, FSinfo host definitions
6180 @comment node-name, next, previous, up
6181 @section @i{FSinfo} filesystems
6182 @cindex FSinfo filesystems
6184 The list of physically attached filesystems follows the machine
6185 attributes. These should define all the filesystems available from this
6186 machine, whether exported or not. In addition to the device name,
6187 filesystems have several attributes, such as filesystem type, mount
6188 options, and @samp{fsck} pass number which are needed to generate
6189 @file{fstab} entries.
6192 filesystem : "fs" @var{<device>} "@{" @i{list(}fs_data@i{)} "@}" ;
6194 fs_data : fs_data_attr "=" @var{<string>}
6199 : "fstype" | "opts" | "passno"
6200 | "freq" | "dumpset" | "log"
6204 Here, @var{<device>} is the device name of the disk (for example,
6205 @file{/dev/dsk/2s0}). The device name is used for building the mount
6206 maps and for the @file{fstab} file. The attributes that can be
6207 specified are shown in the following section.
6209 The @i{FSinfo} configuration file for @code{dylan.doc.ic.ac.uk} is listed below.
6212 host dylan.doc.ic.ac.uk
6220 opts = rw,noquota,grpid
6233 exportfs "dougal eden dylan zebedee brian"
6234 volname /nfs/hp300/local
6245 exportfs "toytown_clients hangers_on"
6246 volname /home/dylan/dk2
6256 exportfs "toytown_clients hangers_on"
6257 volname /home/dylan/dk3
6267 exportfs "toytown_clients hangers_on"
6268 volname /home/dylan/dk5
6274 * FSinfo fstype Option:: FSinfo filesystems fstype.
6275 * FSinfo opts Option:: FSinfo filesystems opts.
6276 * FSinfo passno Option:: FSinfo filesystems passno.
6277 * FSinfo freq Option:: FSinfo filesystems freq.
6278 * FSinfo mount Option:: FSinfo filesystems mount.
6279 * FSinfo dumpset Option:: FSinfo filesystems dumpset.
6280 * FSinfo log Option:: FSinfo filesystems log.
6283 @node FSinfo fstype Option, FSinfo opts Option, , FSinfo filesystems
6284 @comment node-name, next, previous, up
6285 @subsection fstype Option
6286 @cindex FSinfo fstype filesystems option
6287 @cindex fstype, FSinfo filesystems option
6288 @cindex export, FSinfo special fstype
6290 This specifies the type of filesystem being declared and will be placed
6291 into the @file{fstab} file as is. The value of this option will be
6292 handed to @code{mount} as the filesystem type---it should have such
6293 values as @code{4.2}, @code{nfs} or @code{swap}. The value is not
6294 examined for correctness.
6296 There is one special case. If the filesystem type is specified as
6297 @samp{export} then the filesystem information will not be added to the
6298 host's @file{fstab} information, but it will still be visible on the
6299 network. This is useful for defining hosts which contain referenced
6300 volumes but which are not under full control of @i{FSinfo}.
6308 @node FSinfo opts Option, FSinfo passno Option, FSinfo fstype Option, FSinfo filesystems
6309 @comment node-name, next, previous, up
6310 @subsection opts Option
6311 @cindex FSinfo opts filesystems option
6312 @cindex opts, FSinfo filesystems option
6314 This defines any options that should be given to @b{mount}(8) in the
6315 @file{fstab} file. For example:
6318 opts = rw,nosuid,grpid
6321 @node FSinfo passno Option, FSinfo freq Option, FSinfo opts Option, FSinfo filesystems
6322 @comment node-name, next, previous, up
6323 @subsection passno Option
6324 @cindex FSinfo passno filesystems option
6325 @cindex passno, FSinfo filesystems option
6327 This defines the @b{fsck}(8) pass number in which to check the
6328 filesystem. This value will be placed into the @file{fstab} file.
6336 @node FSinfo freq Option, FSinfo mount Option, FSinfo passno Option, FSinfo filesystems
6337 @comment node-name, next, previous, up
6338 @subsection freq Option
6339 @cindex FSinfo freq filesystems option
6340 @cindex freq, FSinfo filesystems option
6342 This defines the interval (in days) between dumps. The value is placed
6343 as is into the @file{fstab} file.
6351 @node FSinfo mount Option, FSinfo dumpset Option, FSinfo freq Option, FSinfo filesystems
6352 @comment node-name, next, previous, up
6353 @subsection mount Option
6354 @cindex FSinfo mount filesystems option
6355 @cindex mount, FSinfo filesystems option
6356 @cindex exportfs, FSinfo mount option
6357 @cindex volname, FSinfo mount option
6358 @cindex sel, FSinfo mount option
6360 This defines the mountpoint at which to place the filesystem. If the
6361 mountpoint of the filesystem is specified as @code{default}, then the
6362 filesystem will be mounted in the automounter's tree under its volume
6363 name and the mount will automatically be inherited by the automounter.
6365 Following the mountpoint, namespace information for the filesystem may
6366 be described. The options that can be given here are @code{exportfs},
6367 @code{volname} and @code{sel}.
6372 mount : "mount" vol_tree ;
6374 vol_tree : @i{list(}vol_tree_attr@i{)} ;
6377 : @var{<string>} "@{" @i{list(}vol_tree_info@i{)} vol_tree "@}" ;
6380 : "exportfs" @var{<export-data>}
6381 | "volname" @var{<volname>}
6382 | "sel" @var{<selector-list>}
6390 exportfs "dylan dougal florence zebedee"
6395 In the above example, the filesystem currently being declared will have
6396 an entry placed into the @file{exports} file allowing the filesystem to
6397 be exported to the machines @code{dylan}, @code{dougal}, @code{florence}
6398 and @code{zebedee}. The volume name by which the filesystem will be
6399 referred to remotely, is @file{/vol/andrew}. By declaring the
6400 mountpoint to be @code{default}, the filesystem will be mounted on the
6401 local machine in the automounter tree, where @i{Amd} will automatically
6402 inherit the mount as @file{/vol/andrew}.@refill
6406 a string defining which machines the filesystem may be exported to.
6407 This is copied, as is, into the @file{exports} file---no sanity checking
6408 is performed on this string.@refill
6411 a string which declares the remote name by which to reference the
6412 filesystem. The string is entered into a dictionary and allows you to
6413 refer to this filesystem in other places by this volume name.@refill
6416 a string which is placed into the automounter maps as a selector for the
6421 @node FSinfo dumpset Option, FSinfo log Option, FSinfo mount Option, FSinfo filesystems
6422 @comment node-name, next, previous, up
6423 @subsection dumpset Option
6424 @cindex FSinfo dumpset filesystems option
6425 @cindex dumpset, FSinfo filesystems option
6427 This provides support for Imperial College's local file backup tools and
6428 is not documented further here.
6430 @node FSinfo log Option, , FSinfo dumpset Option, FSinfo filesystems
6431 @comment node-name, next, previous, up
6432 @subsection log Option
6433 @cindex FSinfo log filesystems option
6434 @cindex log, FSinfo filesystems option
6436 Specifies the log device for the current filesystem. This is ignored if
6437 not required by the particular filesystem type.
6439 @node FSinfo static mounts, FSinfo automount definitions , FSinfo filesystems, FSinfo host definitions
6440 @comment node-name, next, previous, up
6441 @section @i{FSinfo} static mounts
6442 @cindex FSinfo static mounts
6443 @cindex Statically mounts filesystems, FSinfo
6445 Each host may also have a number of statically mounted filesystems. For
6446 example, the host may be a diskless workstation in which case it will
6447 have no @code{fs} declarations. In this case the @code{mount}
6448 declaration is used to determine from where its filesystems will be
6449 mounted. In addition to being added to the @file{fstab} file, this
6450 information can also be used to generate a suitable @file{bootparams}
6454 mount : "mount" @var{<volname>} @i{list(}localinfo@i{)} ;
6456 localinfo : localinfo_attr @var{<string>} ;
6466 The filesystem specified to be mounted will be searched for in the
6467 dictionary of volume names built when scanning the list of hosts'
6470 The attributes have the following semantics:
6472 @item from @var{machine}
6473 mount the filesystem from the machine with the hostname of
6474 @dfn{machine}.@refill
6476 @item as @var{mountpoint}
6477 mount the filesystem locally as the name given, in case this is
6478 different from the advertised volume name of the filesystem.
6480 @item opts @var{options}
6481 native @b{mount}(8) options.
6483 @item fstype @var{type}
6484 type of filesystem to be mounted.
6490 mount /export/exec/hp300/local as /usr/local
6493 If the mountpoint specified is either @file{/} or @file{swap}, the
6494 machine will be considered to be booting off the net and this will be
6495 noted for use in generating a @file{bootparams} file for the host which
6496 owns the filesystems.
6498 @node FSinfo automount definitions, FSinfo Command Line Options, FSinfo static mounts, FSinfo
6499 @comment node-name, next, previous, up
6500 @section Defining an @i{Amd} Mount Map in @i{FSinfo}
6501 @cindex FSinfo automount definitions
6502 @cindex Defining an Amd mount map, FSinfo
6504 The maps used by @i{Amd} can be constructed from @i{FSinfo} by defining
6505 all the automount trees. @i{FSinfo} takes all the definitions found and
6506 builds one map for each top level tree.
6508 The automount tree is usually defined last. A single automount
6509 configuration will usually apply to an entire management domain. One
6510 @code{automount} declaration is needed for each @i{Amd} automount point.
6511 @i{FSinfo} determines whether the automount point is @dfn{direct}
6512 (@pxref{Direct Automount Filesystem}) or @dfn{indirect}
6513 (@pxref{Top-level Filesystem}). Direct automount points are
6514 distinguished by the fact that there is no underlying
6515 @dfn{automount_tree}.@refill
6518 automount : "automount" @i{opt(}auto_opts@i{)} automount_tree ;
6520 auto_opts : "opts" @var{<mount-options>} ;
6523 : @i{list(}automount_attr@i{)}
6527 : @var{<string>} "=" @var{<volname>}
6528 | @var{<string>} "->" @var{<symlink>}
6529 | @var{<string>} "@{" automount_tree "@}"
6533 If @var{<mount-options>} is given, then it is the string to be placed in
6534 the maps for @i{Amd} for the @code{opts} option.
6536 A @dfn{map} is typically a tree of filesystems, for example @file{home}
6537 normally contains a tree of filesystems representing other machines in
6540 A map can either be given as a name representing an already defined
6541 volume name, or it can be a tree. A tree is represented by placing
6542 braces after the name. For example, to define a tree @file{/vol}, the
6543 following map would be defined:
6546 automount /vol @{ @}
6549 Within a tree, the only items that can appear are more maps.
6559 In this case, @i{FSinfo} will look for volumes named @file{/vol/andrew}
6560 and @file{/vol/X11} and a map entry will be generated for each. If the
6561 volumes are defined more than once, then @i{FSinfo} will generate
6562 a series of alternate entries for them in the maps.@refill
6564 Instead of a tree, either a link (@var{name} @code{->}
6565 @var{destination}) or a reference can be specified (@var{name} @code{=}
6566 @var{destination}). A link creates a symbolic link to the string
6567 specified, without further processing the entry. A reference will
6568 examine the destination filesystem and optimize the reference. For
6569 example, to create an entry for @code{njw} in the @file{/homes} map,
6570 either of the two forms can be used:@refill
6574 njw -> /home/dylan/njw
6582 njw = /home/dylan/njw
6586 In the first example, when @file{/homes/njw} is referenced from @i{Amd},
6587 a link will be created leading to @file{/home/dylan/njw} and the
6588 automounter will be referenced a second time to resolve this filename.
6589 The map entry would be:
6592 njw type:=link;fs:=/home/dylan/njw
6595 In the second example, the destination directory is analyzed and found
6596 to be in the filesystem @file{/home/dylan} which has previously been
6597 defined in the maps. Hence the map entry will look like:
6600 njw rhost:=dylan;rfs:=/home/dylan;sublink:=njw
6603 Creating only one symbolic link, and one access to @i{Amd}.
6605 @node FSinfo Command Line Options, FSinfo errors, FSinfo automount definitions, FSinfo
6606 @comment node-name, next, previous, up
6607 @section @i{FSinfo} Command Line Options
6608 @cindex FSinfo command line options
6609 @cindex Command line options, FSinfo
6611 @i{FSinfo} is started from the command line by using the command:
6614 fsinfo [@i{options}] @i{files} ...
6617 The input to @i{FSinfo} is a single set of definitions of machines and
6618 automount maps. If multiple files are given on the command-line, then
6619 the files are concatenated together to form the input source. The files
6620 are passed individually through the C pre-processor before being parsed.
6622 Several options define a prefix for the name of an output file. If the
6623 prefix is not specified no output of that type is produced. The suffix
6624 used will correspond either to the hostname to which a file belongs, or
6625 to the type of output if only one file is produced. Dumpsets and the
6626 @file{bootparams} file are in the latter class. To put the output into
6627 a subdirectory simply put a @file{/} at the end of the prefix, making
6628 sure that the directory has already been made before running
6632 * -a FSinfo Option:: Amd automount directory:
6633 * -b FSinfo Option:: Prefix for bootparams files.
6634 * -d FSinfo Option:: Prefix for dumpset data files.
6635 * -e FSinfo Option:: Prefix for exports files.
6636 * -f FSinfo Option:: Prefix for fstab files.
6637 * -h FSinfo Option:: Local hostname.
6638 * -m FSinfo Option:: Prefix for automount maps.
6639 * -q FSinfo Option:: Ultra quiet mode.
6640 * -v FSinfo Option:: Verbose mode.
6641 * -I FSinfo Option:: Define new #include directory.
6642 * -D-FSinfo Option:: Define macro.
6643 * -U FSinfo Option:: Undefine macro.
6646 @node -a FSinfo Option, -b FSinfo Option, FSinfo Command Line Options, FSinfo Command Line Options
6647 @comment node-name, next, previous, up
6648 @subsection @code{-a} @var{autodir}
6650 Specifies the directory name in which to place the automounter's
6651 mountpoints. This defaults to @file{/a}. Some sites have the autodir set
6652 to be @file{/amd}, and this would be achieved by:
6658 @node -b FSinfo Option, -d FSinfo Option, -a FSinfo Option, FSinfo Command Line Options
6659 @comment node-name, next, previous, up
6660 @subsection @code{-b} @var{bootparams}
6661 @cindex bootparams, FSinfo prefix
6663 This specifies the prefix for the @file{bootparams} filename. If it is
6664 not given, then the file will not be generated. The @file{bootparams}
6665 file will be constructed for the destination machine and will be placed
6666 into a file named @file{bootparams} and prefixed by this string. The
6667 file generated contains a list of entries describing each diskless
6668 client that can boot from the destination machine.
6670 As an example, to create a @file{bootparams} file in the directory
6671 @file{generic}, the following would be used:
6674 fsinfo -b generic/ ...
6677 @node -d FSinfo Option, -e FSinfo Option, -b FSinfo Option, FSinfo Command Line Options
6678 @comment node-name, next, previous, up
6679 @subsection @code{-d} @var{dumpsets}
6680 @cindex dumpset, FSinfo prefix
6682 This specifies the prefix for the @file{dumpsets} file. If it is not
6683 specified, then the file will not be generated. The file will be for
6684 the destination machine and will be placed into a filename
6685 @file{dumpsets}, prefixed by this string. The @file{dumpsets} file is
6686 for use by Imperial College's local backup system.
6688 For example, to create a @file{dumpsets} file in the directory @file{generic},
6689 then you would use the following:
6692 fsinfo -d generic/ ...
6695 @node -e FSinfo Option, -f FSinfo Option, -d FSinfo Option, FSinfo Command Line Options
6696 @comment node-name, next, previous, up
6697 @subsection @code{-e} @var{exportfs}
6698 @cindex exports, FSinfo prefix
6700 Defines the prefix for the @file{exports} files. If it is not given,
6701 then the file will not be generated. For each machine defined in the
6702 configuration files as having disks, an @file{exports} file is
6703 constructed and given a filename determined by the name of the machine,
6704 prefixed with this string. If a machine is defined as diskless, then no
6705 @file{exports} file will be created for it. The files contain entries
6706 for directories on the machine that may be exported to clients.
6708 Example: To create the @file{exports} files for each diskfull machine
6709 and place them into the directory @file{exports}:
6712 fsinfo -e exports/ ...
6715 @node -f FSinfo Option, -h FSinfo Option, -e FSinfo Option, FSinfo Command Line Options
6716 @comment node-name, next, previous, up
6717 @subsection @code{-f} @var{fstab}
6718 @cindex fstab, FSinfo prefix
6720 This defines the prefix for the @file{fstab} files. The files will only
6721 be created if this prefix is defined. For each machine defined in the
6722 configuration files, a @file{fstab} file is created with the filename
6723 determined by prefixing this string with the name of the machine. These
6724 files contain entries for filesystems and partitions to mount at boot
6727 Example, to create the files in the directory @file{fstabs}:
6730 fsinfo -f fstabs/ ...
6733 @node -h FSinfo Option, -m FSinfo Option, -f FSinfo Option, FSinfo Command Line Options
6734 @comment node-name, next, previous, up
6735 @subsection @code{-h} @var{hostname}
6736 @cindex hostname, FSinfo command line option
6738 Defines the hostname of the destination machine to process for. If this
6739 is not specified, it defaults to the local machine name, as returned by
6745 fsinfo -h dylan.doc.ic.ac.uk ...
6748 @node -m FSinfo Option, -q FSinfo Option, -h FSinfo Option, FSinfo Command Line Options
6749 @comment node-name, next, previous, up
6750 @subsection @code{-m} @var{mount-maps}
6751 @cindex maps, FSinfo command line option
6753 Defines the prefix for the automounter files. The maps will only be
6754 produced if this prefix is defined. The mount maps suitable for the
6755 network defined by the configuration files will be placed into files
6756 with names calculated by prefixing this string to the name of each map.
6758 For example, to create the automounter maps and place them in the
6759 directory @file{automaps}:
6762 fsinfo -m automaps/ ...
6765 @node -q FSinfo Option, -v FSinfo Option, -m FSinfo Option, FSinfo Command Line Options
6766 @comment node-name, next, previous, up
6767 @subsection @code{-q}
6768 @cindex quiet, FSinfo command line option
6770 Selects quiet mode. @i{FSinfo} suppress the ``running commentary'' and
6771 only outputs any error messages which are generated.
6773 @node -v FSinfo Option, -D-FSinfo Option, -q FSinfo Option, FSinfo Command Line Options
6774 @comment node-name, next, previous, up
6775 @subsection @code{-v}
6776 @cindex verbose, FSinfo command line option
6778 Selects verbose mode. When this is activated, the program will display
6779 more messages, and display all the information discovered when
6780 performing the semantic analysis phase. Each verbose message is output
6781 to @file{stdout} on a line starting with a @samp{#} character.
6783 @node -D-FSinfo Option, -I FSinfo Option, -v FSinfo Option, FSinfo Command Line Options
6784 @comment node-name, next, previous, up
6785 @subsection @code{-D} @var{name}@i{[=defn]}
6787 Defines a symbol @dfn{name} for the preprocessor when reading the
6788 configuration files. Equivalent to @code{#define} directive.
6790 @node -I FSinfo Option, -U FSinfo Option, -D-FSinfo Option, FSinfo Command Line Options
6791 @comment node-name, next, previous, up
6792 @subsection @code{-I} @var{directory}
6794 This option is passed into the preprocessor for the configuration files.
6795 It specifies directories in which to find include files
6797 @node -U FSinfo Option, , -I FSinfo Option, FSinfo Command Line Options
6798 @comment node-name, next, previous, up
6799 @subsection @code{-U} @var{name}
6801 Removes any initial definition of the symbol @dfn{name}. Inverse of the
6804 @node FSinfo errors, , FSinfo Command Line Options, FSinfo
6805 @comment node-name, next, previous, up
6806 @section Errors produced by @i{FSinfo}
6807 @cindex FSinfo error messages
6809 The following table documents the errors and warnings which @i{FSinfo} may produce.
6814 Occurs if an unescaped newline is found in a quoted string.
6816 @item ambiguous mount: @var{volume} is a replicated filesystem
6817 If several filesystems are declared as having the same volume name, they
6818 will be considered replicated filesystems. To mount a replicated
6819 filesystem statically, a specific host will need to be named, to say
6820 which particular copy to try and mount, else this error will
6823 @item can't open @var{filename} for writing
6824 Occurs if any errors are encountered when opening an output file.
6826 @item cannot determine localname since volname @var{volume} is not uniquely defined
6827 If a volume is replicated and an attempt is made to mount the filesystem
6828 statically without specifying a local mountpoint, @i{FSinfo} cannot
6829 calculate a mountpoint, as the desired pathname would be
6832 @item @var{device} has duplicate exportfs data
6833 Produced if the @samp{exportfs} option is used multiple times within the
6834 same branch of a filesystem definition. For example, if you attempt to
6835 set the @samp{exportfs} data at different levels of the mountpoint
6838 @item dump frequency for @var{host}:@var{device} is non-zero
6839 Occurs if @var{device} has its @samp{fstype} declared to be @samp{swap}
6840 or @samp{export} and the @samp{dump} option is set to a value greater
6841 than zero. Swap devices should not be dumped.
6843 @item duplicate host @var{hostname}!
6844 If a host has more than one definition.
6846 @item end of file within comment
6847 A comment was unterminated before the end of one of the configuration
6850 @item @var{filename}: cannot open for reading
6851 If a file specified on the command line as containing configuration data
6852 could not be opened.
6854 @item @var{filesystem} has a volname but no exportfs data
6855 Occurs when a volume name is declared for a file system, but the string
6856 specifying what machines the filesystem can be exported to is
6859 @item fs field "@var{field-name}" already set
6860 Occurs when multiple definitions are given for one of the attributes of a
6863 @item host field "@var{field-name}" already set
6864 If duplicate definitions are given for any of the fields with a host
6867 @item @var{host}:@var{device} has more than one mount point
6868 Occurs if the mount option for a host's filesystem specifies multiple
6869 trees at which to place the mountpoint.
6871 @item @var{host}:@var{device} has no mount point
6872 Occurs if the @samp{mount} option is not specified for a host's
6875 @item @var{host}:@var{device} needs field "@var{field-name}"
6876 Occurs when a filesystem is missing a required field. @var{field-name} could
6877 be one of @samp{fstype}, @samp{opts}, @samp{passno} or
6880 @item @var{host}:mount field specified for swap partition
6881 Occurs if a mountpoint is given for a filesystem whose type is declared
6884 @item malformed IP dotted quad: @var{address}
6885 If the Internet address of an interface is incorrectly specified. An
6886 Internet address definition is handled to @b{inet_addr}(3N) to see if it
6887 can cope. If not, then this message will be displayed.
6889 @item malformed netmask: @var{netmask}
6890 If the netmask cannot be decoded as though it were a hexadecimal number,
6891 then this message will be displayed. It will typically be caused by
6892 incorrect characters in the @var{netmask} value.
6894 @item mount field "@var{field-name}" already set
6895 Occurs when a static mount has multiple definitions of the same field.
6897 @item mount tree field "@var{field-name}" already set
6898 Occurs when the @var{field-name} is defined more than once during the
6899 definition of a filesystems mountpoint.
6901 @item netif field @var{field-name} already set
6902 Occurs if you attempt to define an attribute of an interface more than
6905 @item network booting requires both root and swap areas
6906 Occurs if a machine has mount declarations for either the root partition
6907 or the swap area, but not both. You cannot define a machine to only
6908 partially boot via the network.
6910 @item no disk mounts on @var{hostname}
6911 If there are no static mounts, nor local disk mounts specified for a
6912 machine, this message will be displayed.
6914 @item no volname given for @var{host}:@var{device}
6915 Occurs when a filesystem is defined to be mounted on @file{default}, but
6916 no volume name is given for the file system, then the mountpoint cannot
6919 @item not allowed '/' in a directory name
6920 Occurs when a pathname with multiple directory elements is specified as
6921 the name for an automounter tree. A tree should only have one name at
6924 @item pass number for @var{host}:@var{device} is non-zero
6925 Occurs if @var{device} has its @samp{fstype} declared to be @samp{swap}
6926 or @samp{export} and the @b{fsck}(8) pass number is set. Swap devices
6927 should not be fsck'd. @xref{FSinfo fstype Option}.
6929 @item sub-directory @var{directory} of @var{directory-tree} starts with '/'
6930 Within the filesystem specification for a host, if an element
6931 @var{directory} of the mountpoint begins with a @samp{/} and it is not
6932 the start of the tree.
6934 @item sub-directory of @var{directory-tree} is named "default"
6935 @samp{default} is a keyword used to specify if a mountpoint should be
6936 automatically calculated by @i{FSinfo}. If you attempt to specify a
6937 directory name as this, it will use the filename of @file{default} but
6938 will produce this warning.
6940 @item unknown \ sequence
6941 Occurs if an unknown escape sequence is found inside a string. Within a
6942 string, you can give the standard C escape sequences for strings, such
6943 as newlines and tab characters.
6945 @item unknown directory attribute
6946 If an unknown keyword is found while reading the definition of a host's
6947 filesystem mount option.
6949 @item unknown filesystem attribute
6950 Occurs if an unrecognized keyword is used when defining a host's
6953 @item unknown host attribute
6954 Occurs if an unrecognized keyword is used when defining a host.
6956 @item unknown mount attribute
6957 Occurs if an unrecognized keyword is found while parsing the list of
6960 @item unknown volname @var{volume} automounted @i{[} on @i{name} @i{]}
6961 Occurs if @var{volume} is used in a definition of an automount map but the volume
6962 name has not been declared during the host filesystem definitions.
6964 @item volname @var{volume} is unknown
6965 Occurs if an attempt is made to mount or reference a volume name which
6966 has not been declared during the host filesystem definitions.
6968 @item volname @var{volume} not exported from @var{machine}
6969 Occurs if you attempt to mount the volume @var{volume} from a machine
6970 which has not declared itself to have such a filesystem
6975 @c ################################################################
6976 @node Hlfsd, Assorted Tools, FSinfo, Top
6977 @comment node-name, next, previous, up
6980 @cindex Home-Link Filesystem
6982 @i{Hlfsd} is a daemon which implements a filesystem containing a
6983 symbolic link to subdirectory within a user's home directory, depending
6984 on the user which accessed that link. It was primarily designed to
6985 redirect incoming mail to users' home directories, so that it can be read
6986 from anywhere. It was designed and implemented by
6987 @uref{http://www.cs.sunysb.edu/~ezk,Erez Zadok} and
6988 @email{dupuy AT cs.columbia.edu,Alexander Dupuy}, at the
6989 @uref{http://www.cs.columbia.edu/,Computer Science Department} of
6990 @uref{http://www.columbia.edu/,Columbia University}. A
6991 @uref{http://www.fsl.cs.sunysb.edu/docs/hlfsd/hlfsd.html,paper}
6992 on @i{Hlfsd} was presented at the Usenix LISA VII conference in 1993.
6994 @i{Hlfsd} operates by mounting itself as an NFS server for the directory
6995 containing @i{linkname}, which defaults to @file{/hlfs/home}. Lookups
6996 within that directory are handled by @i{Hlfsd}, which uses the
6997 password map to determine how to resolve the lookup. The directory will
6998 be created if it doesn't already exist. The symbolic link will be to
6999 the accessing user's home directory, with @i{subdir} appended to it. If
7000 not specified, @i{subdir} defaults to @file{.hlfsdir}. This directory
7001 will also be created if it does not already exist.
7003 A @samp{SIGTERM} sent to @i{Hlfsd} will cause it to shutdown. A
7004 @samp{SIGHUP} will flush the internal caches, and reload the password
7005 map. It will also close and reopen the log file, to enable the original
7006 log file to be removed or rotated. A @samp{SIGUSR1} will cause it to
7007 dump its internal table of user IDs and home directories to the file
7008 @file{/tmp/hlfsddump}.
7011 * Introduction to Hlfsd::
7012 * Background to Mail Delivery::
7016 @c ================================================================
7017 @node Introduction to Hlfsd, Background to Mail Delivery, Hlfsd, Hlfsd
7018 @comment node-name, next, previous, up
7019 @section Introduction to Hlfsd
7020 @cindex Introduction to Hlfsd
7021 @cindex Hlfsd; introduction
7023 Electronic mail has become one of the major applications for many
7024 computer networks, and use of this service is expected to increase over
7025 time, as networks proliferate and become faster. Providing a convenient
7026 environment for users to read, compose, and send electronic mail has
7027 become a requirement for systems administrators (SAs).
7029 Widely used methods for handling mail usually require users to be logged
7030 into a designated ``home'' machine, where their mailbox files reside.
7031 Only on that one machine can they read newly arrived mail. Since users
7032 have to be logged into that system to read their mail, they often find
7033 it convenient to run all of their other processes on that system as
7034 well, including memory and CPU-intensive jobs. For example, in our
7035 department, we have allocated and configured several multi-processor
7036 servers to handle such demanding CPU/memory applications, but these were
7037 underutilized, in large part due to the inconvenience of not being able
7038 to read mail on those machines. (No home directories were located on
7039 these designated CPU-servers, since we did not want NFS service for
7040 users' home directories to have to compete with CPU-intensive jobs. At the
7041 same time, we discouraged users from running demanding applications on
7042 their home machines.)
7044 Many different solutions have been proposed to allow users to read their
7045 mail on any host. However, all of these solutions fail in one or more
7051 they introduce new single points of failure
7054 they require using different mail transfer agents (MTAs) or user agents
7058 they do not solve the problem for all cases, i.e. the solution is only
7059 partially successful for a particular environment.
7063 We have designed a simple filesystem, called the @dfn{Home-Link File
7064 System}, to provide the ability to deliver mail to users' home
7065 directories, without modification to mail-related applications. We have
7066 endeavored to make it as stable as possible. Of great importance to us
7067 was to make sure the HLFS daemon, @file{hlfsd} , would not hang under
7068 any circumstances, and would take the next-best action when faced with
7069 problems. Compared to alternative methods, @i{Hlfsd} is a stable, more
7070 general solution, and easier to install/use. In fact, in some ways, we
7071 have even managed to improve the reliability and security of mail
7074 Our server implements a small filesystem containing a symbolic link
7075 to a subdirectory of the invoking user's home directory, and named symbolic
7076 links to users' mailbox files.
7078 The @i{Hlfsd} server finds out the @var{uid} of the process that is
7079 accessing its mount point, and resolves the pathname component @samp{home} as a
7080 symbolic link to a subdirectory within the home directory given by the
7081 @var{uid}'s entry in the password file. If the @var{gid} of the process
7082 that attempts to access a mailbox file is a special one (called
7083 HLFS_GID), then the server maps the name of the @emph{next} pathname
7084 component directly to the user's mailbox. This is necessary so that
7085 access to a mailbox file by users other than the owner can succeed. The
7086 server has safety features in case of failures such as hung filesystems
7087 or home directory filesystems that are inaccessible or full.
7089 On most of our machines, mail gets delivered to the directory
7090 @file{/var/spool/mail}. Many programs, including UAs, depend on that
7091 path. @i{Hlfsd} creates a directory @file{/mail}, and mounts itself on
7092 top of that directory. @i{Hlfsd} implements the path name component
7093 called @samp{home}, pointing to a subdirectory of the user's home directory.
7094 We have made @file{/var/spool/mail} a symbolic link to
7095 @file{/mail/home}, so that accessing @file{/var/spool/mail} actually
7096 causes access to a subdirectory within a user's home directory.
7098 The following table shows an example of how resolving the pathname
7099 @file{/var/mail/@i{NAME}} to @file{/users/ezk/.mailspool/@i{NAME}} proceeds.
7101 @multitable {Resolving Component} {Pathname left to resolve} {Value if symbolic link}
7103 @item @b{Resolving Component}
7104 @tab @b{Pathname left to resolve}
7105 @tab @b{Value if symbolic link}
7108 @tab @t{var/mail/}@i{NAME}
7111 @tab @t{mail/}@i{NAME}
7114 @tab @t{/mail/home/}@i{NAME}
7115 @tab @t{mail}@@ -> @t{/mail/home}
7118 @tab @t{mail/home/}@i{NAME}
7121 @tab @t{home/}@i{NAME}
7125 @tab @t{home}@@ -> @t{/users/ezk/.mailspool}
7128 @tab @t{users/ezk/.mailspool/}@i{NAME}
7131 @tab @t{ezk/.mailspool/}@i{NAME}
7134 @tab @t{.mailspool/}@i{NAME}
7136 @item @t{.mailspool/}
7143 @c ================================================================
7144 @node Background to Mail Delivery, Using Hlfsd, Introduction to Hlfsd, Hlfsd
7145 @comment node-name, next, previous, up
7146 @section Background to Mail Delivery
7147 @cindex Background to Mail Delivery
7148 @cindex Hlfsd; background
7150 This section provides an in-depth discussion of why available methods
7151 for delivering mail to home directories are not as good as the one used
7155 * Single-Host Mail Spool Directory::
7156 * Centralized Mail Spool Directory::
7157 * Distributed Mail Spool Service::
7158 * Why Deliver Into the Home Directory?::
7161 @c ----------------------------------------------------------------
7162 @node Single-Host Mail Spool Directory, Centralized Mail Spool Directory, Background to Mail Delivery, Background to Mail Delivery
7163 @comment node-name, next, previous, up
7164 @subsection Single-Host Mail Spool Directory
7165 @cindex Single-Host Mail Spool Directory
7167 The most common method for mail delivery is for mail to be appended to a
7168 mailbox file in a standard spool directory on the designated ``mail
7169 home'' machine of the user. The greatest advantage of this method is
7170 that it is the default method most vendors provide with their systems,
7171 thus very little (if any) configuration is required on the SA's part.
7172 All they need to set up are mail aliases directing mail to the host on
7173 which the user's mailbox file is assigned. (Otherwise, mail is
7174 delivered locally, and users find mailboxes on many machines.)
7176 As users become more sophisticated, and aided by windowing systems, they
7177 find themselves logging in on multiple hosts at once, performing several
7178 tasks concurrently. They ask to be able to read their mail on any host
7179 on the network, not just the one designated as their ``mail home''.
7181 @c ----------------------------------------------------------------
7182 @node Centralized Mail Spool Directory, Distributed Mail Spool Service, Single-Host Mail Spool Directory, Background to Mail Delivery
7183 @comment node-name, next, previous, up
7184 @subsection Centralized Mail Spool Directory
7185 @cindex Centralized Mail Spool Directory
7187 A popular method for providing mail readability from any host is to have
7188 all mail delivered to a mail spool directory on a designated
7189 ``mail-server'' which is exported via NFS to all of the hosts on the
7190 network. Configuring such a system is relatively easy. On most
7191 systems, the bulk of the work is a one-time addition to one or two
7192 configuration files in @file{/etc}. The file-server's spool directory
7193 is then hard-mounted across every machine on the local network. In
7194 small environments with only a handful of hosts this can be an
7195 acceptable solution. In our department, with a couple of hundred active
7196 hosts and thousands of mail messages processed daily, this was deemed
7197 completely unacceptable, as it introduced several types of problems:
7201 @item Scalability and Performance
7203 As more and more machines get added to the network, more mail traffic
7204 has to go over NFS to and from the mail-server. Users like to run
7205 mail-watchers, and read their mail often. The stress on the shared
7206 infrastructure increases with every user and host added; loads on the
7207 mail server would most certainly be high since all mail delivery goes
7208 through that one machine.@footnote{ Delivery via NFS-mounted filesystems
7209 may require usage of @samp{rpc.lockd} and @samp{rpc.statd} to provide
7210 distributed file-locking, both of which are widely regarded as unstable
7211 and unreliable. Furthermore, this will degrade performance, as local
7212 processes as well as remote @samp{nfsd} processes are kept busy.} This
7213 leads to lower reliability and performance. To reduce the number of
7214 concurrent connections between clients and the server host, some SAs
7215 have resorted to automounting the mail-spool directory. But this
7216 solution only makes things worse: since users often run mail watchers,
7217 and many popular applications such as @samp{trn}, @samp{emacs},
7218 @samp{csh} or @samp{ksh} check periodically for new mail, the
7219 automounted directory would be effectively permanently mounted. If it
7220 gets unmounted automatically by the automounter program, it is most
7221 likely to get mounted shortly afterwards, consuming more I/O resources
7222 by the constant cycle of mount and umount calls.
7226 The mail-server host and its network connectivity must be very reliable.
7227 Worse, since the spool directory has to be hard-mounted,@footnote{No SA
7228 in their right minds would soft-mount read/write partitions --- the
7229 chances for data loss are too great.} many processes which access the
7230 spool directory (various shells, @samp{login}, @samp{emacs}, etc.)
7231 would be hung as long as connectivity to the mail-server is severed. To
7232 improve reliability, SAs may choose to backup the mail-server's spool
7233 partition several times a day. This may make things worse since reading
7234 or delivering mail while backups are in progress may cause backups to be
7235 inconsistent; more backups consume more backup-media resources, and
7236 increase the load on the mail-server host.
7240 @c ----------------------------------------------------------------
7241 @node Distributed Mail Spool Service, Why Deliver Into the Home Directory?, Centralized Mail Spool Directory, Background to Mail Delivery
7242 @comment node-name, next, previous, up
7243 @subsection Distributed Mail Spool Service
7244 @cindex Distributed Mail Spool Service
7246 Despite the existence of a few systems that support delivery to users'
7247 home directories, mail delivery to home directories hasn't caught on.
7248 We believe the main reason is that there are too many programs that
7249 ``know'' where mailbox files reside. Besides the obvious (the delivery
7250 program @file{/bin/mail} and mail readers like @file{/usr/ucb/Mail},
7251 @samp{mush}, @samp{mm}, etc.), other programs that know mailbox location
7252 are login, from, almost every shell, @samp{xbiff}, @samp{xmailbox}, and
7253 even some programs not directly related to mail, such as @samp{emacs}
7254 and @samp{trn}. Although some of these programs can be configured to
7255 look in different directories with the use of environment variables and
7256 other resources, many of them cannot. The overall porting work is
7259 Other methods that have yet to catch on require the use of a special
7260 mail-reading server, such as IMAP or POP. The main disadvantage of
7261 these systems is that UAs need to be modified to use these services ---
7262 a long and involved task. That is why they are not popular at this
7265 Several other ideas have been proposed and even used in various
7266 environments. None of them is robust. They are mostly very
7267 specialized, inflexible, and do not extend to the general case. Some of
7268 the ideas are plain bad, potentially leading to lost or corrupt mail:
7274 Using an automounter such as @i{Amd} to provide a set of symbolic links
7275 from the normal spool directory to user home directories is not
7276 sufficient. UAs rename, unlink, and recreate the mailbox as a regular
7277 file, therefore it must be a real file, not a symbolic link.
7278 Furthermore, it must reside in a real directory which is writable by the
7279 UAs and MTAs. This method may also require populating
7280 @file{/var/spool/mail} with symbolic links and making sure they are
7281 updated. Making @i{Amd} manage that directory directly fails, since
7282 many various lock files need to be managed as well. Also, @i{Amd} does
7283 not provide all of the NFS operations which are required to write mail
7284 such as write, create, remove, and unlink.
7288 Setting this variable to an automounted directory pointing to the user's
7289 mail spool host only solves the problem for those programs which know
7290 and use @code{$MAIL}. Many programs don't, therefore this solution is partial
7291 and of limited flexibility. Also, it requires the SAs or the users to
7292 set it themselves --- an added level of inconvenience and possible
7297 Using a different mail delivery agent could be the solution. One such
7298 example is @samp{hdmail}. However, @samp{hdmail} still requires
7299 modifying all UAs, the MTA's configuration, installing new daemons, and
7300 changing login scripts. This makes the system less upgradable or
7301 compatible with others, and adds one more complicated system for SAs to
7302 deal with. It is not a complete solution because it still requires each
7303 user have their @code{$MAIL} variable setup correctly, and that every program
7308 @c ----------------------------------------------------------------
7309 @node Why Deliver Into the Home Directory?, , Distributed Mail Spool Service, Background to Mail Delivery
7310 @comment node-name, next, previous, up
7311 @subsection Why Deliver Into the Home Directory?
7312 @cindex Why Deliver Into the Home Directory?
7313 @cindex Hlfsd; Why Deliver Into the Home Directory?
7315 There are several major reasons why SAs might want to deliver mail
7316 directly into the users' home directories:
7322 Many mail readers need to move mail from the spool directory to the
7323 user's home directory. It speeds up this operation if the two are on
7324 the same filesystem. If for some reason the user's home directory is
7325 inaccessible, it isn't that useful to be able to read mail, since there
7326 is no place to move it to. In some cases, trying to move mail to a
7327 non-existent or hung filesystem may result in mail loss.
7331 Having all mail spool directories spread among the many more filesystems
7332 minimizes the chances that complete environments will grind to a halt
7333 when a single server is down. It does increase the chance that there
7334 will be someone who is not able to read their mail when a machine is
7335 down, but that is usually preferred to having no one be able to read
7336 their mail because a centralized mail server is down. The problem of
7337 losing some mail due to the (presumably) higher chances that a user's
7338 machine is down is minimized in HLFS.
7342 Delivering mail to users' home directories has another advantage ---
7343 enhanced security and privacy. Since a shared system mail spool
7344 directory has to be world-readable and searchable, any user can see
7345 whether other users have mail, when they last received new mail, or when
7346 they last read their mail. Programs such as @samp{finger} display this
7347 information, which some consider an infringement of privacy. While it
7348 is possible to disable this feature of @samp{finger} so that remote
7349 users cannot see a mailbox file's status, this doesn't prevent local
7350 users from getting the information. Furthermore, there are more
7351 programs which make use of this information. In shared environments,
7352 disabling such programs has to be done on a system-wide basis, but with
7353 mail delivered to users' home directories, users less concerned with
7354 privacy who do want to let others know when they last received or read
7355 mail can easily do so using file protection bits.
7357 @c Lastly, on systems that do not export their NFS filesystem with
7358 @c @t{anon=0}, superusers are less likely to snoop around others' mail, as
7359 @c they become ``nobodies'' across NFS.
7363 In summary, delivering mail to home directories provides users the
7364 functionality sought, and also avoids most of the problems just
7367 @c ================================================================
7368 @node Using Hlfsd, , Background to Mail Delivery, Hlfsd
7369 @comment node-name, next, previous, up
7370 @section Using Hlfsd
7372 @cindex Hlfsd; using
7375 * Controlling Hlfsd::
7380 @c ----------------------------------------------------------------
7381 @node Controlling Hlfsd, Hlfsd Options, Using Hlfsd, Using Hlfsd
7382 @comment node-name, next, previous, up
7383 @subsection Controlling Hlfsd
7384 @cindex Controlling Hlfsd
7385 @cindex Hlfsd; controlling
7388 Much the same way @i{Amd} is controlled by @file{ctl-amd}, so does
7389 @i{Hlfsd} get controlled by the @file{ctl-hlfsd} script:
7393 @item ctl-hlfsd start
7394 Start a new @i{Hlfsd}.
7396 @item ctl-hlfsd stop
7397 Stop a running @i{Hlfsd}.
7399 @item ctl-hlfsd restart
7400 Stop a running @i{Hlfsd}, wait for 10 seconds, and then start a new
7401 one. It is hoped that within 10 seconds, the previously running
7402 @i{Hlfsd} terminate properly; otherwise, starting a second one could
7403 cause system lockup.
7407 For example, on our systems, we start @i{Hlfsd} within @file{ctl-hlfsd}
7408 as follows on Solaris 2 systems:
7411 hlfsd -a /var/alt_mail -x all -l /var/log/hlfsd /mail/home .mailspool
7414 The directory @file{/var/alt_mail} is a directory in the root partition
7415 where alternate mail will be delivered into, when it cannot be delivered
7416 into the user's home directory.
7418 Normal mail gets delivered into @file{/var/mail}, but on our systems,
7419 that is a symbolic link to @file{/mail/home}. @file{/mail} is managed
7420 by @i{Hlfsd}, which creates a dynamic symlink named @samp{home},
7421 pointing to the subdirectory @file{.mailspool} @emph{within} the
7422 accessing user's home directory. This results in mail which normally
7423 should go to @file{/var/mail/@code{$USER}}, to go to
7424 @file{@code{$HOME}/.mailspool/@code{$USER}}.
7426 @i{Hlfsd} does not create the @file{/var/mail} symlink. This needs to
7427 be created (manually) once on each host, by the system administrators,
7431 mv /var/mail /var/alt_mail
7432 ln -s /mail/home /var/mail
7435 @i{Hlfsd} also responds to the following signals:
7437 A @samp{SIGHUP} signal sent to @i{Hlfsd} will force it to reload the
7438 password map immediately.
7440 A @samp{SIGUSR1} signal sent to @i{Hlfsd} will cause it to dump its
7441 internal password map to the file @file{/usr/tmp/hlfsd.dump.XXXXXX},
7442 where @samp{XXXXXX} will be replaced by a random string generated by
7443 @b{mktemp}(3) or (the more secure) @b{mkstemp}(3).
7445 @c ----------------------------------------------------------------
7446 @node Hlfsd Options, Hlfsd Files, Controlling Hlfsd, Using Hlfsd
7447 @comment node-name, next, previous, up
7448 @subsection Hlfsd Options
7449 @cindex Hlfsd Options
7450 @cindex Hlfsd; Options
7454 @item -a @var{alt_dir}
7455 Alternate directory. The name of the directory to which the symbolic
7456 link returned by @i{Hlfsd} will point, if it cannot access the home
7457 directory of the user. This defaults to @file{/var/hlfs}. This
7458 directory will be created if it doesn't exist. It is expected that
7459 either users will read these files, or the system administrators will
7460 run a script to resend this ``lost mail'' to its owner.
7462 @item -c @var{cache-interval}
7463 Caching interval. @i{Hlfsd} will cache the validity of home directories
7464 for this interval, in seconds. Entries which have been verified within
7465 the last @var{cache-interval} seconds will not be verified again, since
7466 the operation could be expensive, and the entries are most likely still
7467 valid. After the interval has expired, @i{Hlfsd} will re-verify the
7468 validity of the user's home directory, and reset the cache time-counter.
7469 The default value for @var{cache-interval} is 300 seconds (5 minutes).
7472 Force fast startup. This option tells @i{Hlfsd} to skip startup-time
7473 consistency checks such as existence of mount directory, alternate spool
7474 directory, symlink to be hidden under the mount directory, their
7475 permissions and validity.
7477 @item -g @var{group}
7478 Set the special group HLFS_GID to @var{group}. Programs such as
7479 @file{/usr/ucb/from} or @file{/usr/sbin/in.comsat}, which access the
7480 mailboxes of other users, must be setgid @samp{HLFS_GID} to work properly. The
7481 default group is @samp{hlfs}. If no group is provided, and there is no
7482 group @samp{hlfs}, this feature is disabled.
7485 Help. Print a brief help message, and exit.
7487 @item -i @var{reload-interval}
7488 Map-reloading interval. Each @var{reload-interval} seconds, @i{Hlfsd}
7489 will reload the password map. @i{Hlfsd} needs the password map for the
7490 UIDs and home directory pathnames. @i{Hlfsd} schedules a @samp{SIGALRM} to
7491 reload the password maps. A @samp{SIGHUP} sent to @i{Hlfsd} will force it to
7492 reload the maps immediately. The default value for
7493 @var{reload-interval} is 900 seconds (15 minutes.)
7495 @item -l @var{logfile}
7496 Specify a log file to which @i{Hlfsd} will record events. If
7497 @var{logfile} is the string @samp{syslog} then the log messages will be
7498 sent to the system log daemon by @b{syslog}(3), using the @samp{LOG_DAEMON}
7499 facility. This is also the default.
7502 No verify. @i{Hlfsd} will not verify the validity of the symbolic link
7503 it will be returning, or that the user's home directory contains
7504 sufficient disk-space for spooling. This can speed up @i{Hlfsd} at the
7505 cost of possibly returning symbolic links to home directories which are
7506 not currently accessible or are full. By default, @i{Hlfsd} validates
7507 the symbolic-link in the background. The @code{-n} option overrides the
7508 meaning of the @code{-c} option, since no caching is necessary.
7510 @item -o @var{mount-options}
7511 Mount options which @i{Hlfsd} will use to mount itself on top of
7512 @var{dirname}. By default, @var{mount-options} is set to @samp{ro}. If
7513 the system supports symbolic-link caching, default options are set
7514 to @samp{ro,nocache}.
7517 Print PID. Outputs the process-id of @i{Hlfsd} to standard output where
7518 it can be saved into a file.
7521 Version. Displays version information to standard error.
7523 @item -x @var{log-options}
7524 Specify run-time logging options. The options are a comma separated
7525 list chosen from: @samp{fatal}, @samp{error}, @samp{user}, @samp{warn}, @samp{info}, @samp{map}, @samp{stats}, @samp{all}.
7528 Force @i{Hlfsd} to run on systems that cannot turn off the NFS
7529 attribute-cache. Use of this option on those systems is discouraged, as
7530 it may result in loss or misdelivery of mail. The option is ignored on
7531 systems that can turn off the attribute-cache.
7533 @item -D @var{log-options}
7534 Select from a variety of debugging options. Prefixing an option with
7535 the string @samp{no} reverses the effect of that option. Options are
7536 cumulative. The most useful option is @samp{all}. Since this option is
7537 only used for debugging other options are not documented here. A fuller
7538 description is available in the program source.
7540 @item -P @var{password-file}
7541 Read the user-name, user-id, and home directory information from the
7542 file @var{password-file}. Normally, @i{Hlfsd} will use @b{getpwent}(3)
7543 to read the password database. This option allows you to override the
7544 default database, and is useful if you want to map users' mail files to
7545 a directory other than their home directory. Only the username, uid,
7546 and home-directory fields of the file @var{password-file} are read and
7547 checked. All other fields are ignored. The file @var{password-file}
7548 must otherwise be compliant with Unix Version 7 colon-delimited format
7553 @c ----------------------------------------------------------------
7554 @node Hlfsd Files, , Hlfsd Options, Using Hlfsd
7555 @comment node-name, next, previous, up
7556 @subsection Hlfsd Files
7558 @cindex Hlfsd; Files
7560 The following files are used by @i{Hlfsd}:
7565 directory under which @i{Hlfsd} mounts itself and manages the symbolic
7569 default sub-directory in the user's home directory, to which the
7570 @file{home} symbolic link returned by @i{Hlfsd} points.
7573 directory to which @file{home} symbolic link returned by @i{Hlfsd}
7574 points if it is unable to verify the that user's home directory is
7577 @item /usr/tmp/hlfsd.dump.XXXXXX
7578 file to which @i{Hlfsd} will dump its internal password map when it
7579 receives the @samp{SIGUSR1} signal. @samp{XXXXXX} will be replaced by
7580 a random string generated by @b{mktemp}(3) or (the more secure)
7585 For discussion on other files used by @i{Hlfsd}, see @xref{lostaltmail}, and
7586 @ref{lostaltmail.conf-sample}.
7588 @c ################################################################
7589 @node Assorted Tools, Examples, Hlfsd, Top
7590 @comment node-name, next, previous, up
7591 @chapter Assorted Tools
7592 @cindex Assorted Tools
7594 The following are additional utilities and scripts included with
7595 am-utils, and get installed.
7609 * lostaltmail.conf-sample::
7618 @c ----------------------------------------------------------------
7619 @node am-eject, amd.conf-sample, Assorted Tools, Assorted Tools
7620 @comment node-name, next, previous, up
7624 A shell script unmounts a floppy or CD-ROM that is automounted, and
7625 then attempts to eject the removable device.
7627 @c ----------------------------------------------------------------
7628 @node amd.conf-sample, amd2ldif, am-eject, Assorted Tools
7629 @comment node-name, next, previous, up
7630 @section amd.conf-sample
7631 @pindex amd.conf-sample
7633 A sample @i{Amd} configuration file. @xref{Amd Configuration File}.
7635 @c ----------------------------------------------------------------
7636 @node amd2ldif, amd2sun, amd.conf-sample, Assorted Tools
7637 @comment node-name, next, previous, up
7641 A script to convert @i{Amd} maps to LDAP input files. Use it as follows:
7644 amd2ldif @i{mapname} @i{base} < @i{amd.mapfile} > @i{mapfile.ldif}
7647 @c ----------------------------------------------------------------
7648 @node amd2sun, automount2amd, amd2ldif, Assorted Tools
7649 @comment node-name, next, previous, up
7653 A script to convert @i{Amd} maps to Sun Automounter maps. Use it as
7657 amd2sun < @i{amd.mapfile} > @i{auto_mapfile}
7660 @c ----------------------------------------------------------------
7661 @node automount2amd, ctl-amd, amd2sun, Assorted Tools
7662 @comment node-name, next, previous, up
7663 @section automount2amd
7664 @pindex automount2amd
7666 A script to convert old Sun Automounter maps to @i{Amd} maps.
7668 Say you have the Sun automount file @i{auto.foo}, with these two lines:
7671 moon -ro,intr server:/proj/images
7675 automount2amd auto.foo > amd.foo
7678 will produce the @i{Amd} map @i{amd.foo} with this content:
7681 # generated by automount2amd on Sat Aug 14 17:59:32 US/Eastern 1999
7684 type:=nfs;opts:=rw,grpid,nosuid,utimeout=600
7687 host==earth;type:=link;fs:=/home \\
7688 rhost:=earth;rfs:=/home
7691 -addopts:=ro,intr \\
7692 host==server;type:=link;fs:=/proj/images \\
7693 rhost:=server;rfs:=/proj/images
7696 This perl script will use the following @i{/default} entry
7698 type:=nfs;opts:=rw,grpid,nosuid,utimeout=600
7700 If you wish to override that, define the @b{$DEFAULTS} environment
7701 variable, or modify the script.
7703 If you wish to generate Amd maps using the @i{hostd} (@pxref{hostd
7704 Selector Variable}) @i{Amd} map syntax, then define the environment
7705 variable @b{$DOMAIN} or modify the script.
7707 Note that automount2amd does not understand the syntax in newer Sun
7708 Automount maps, those used with autofs.
7710 @c ----------------------------------------------------------------
7711 @node ctl-amd, ctl-hlfsd, automount2amd, Assorted Tools
7712 @comment node-name, next, previous, up
7716 A script to start, stop, or restart @i{Amd}. Use it as follows:
7720 Start a new @i{Amd} process.
7722 Stop the running @i{Amd}.
7723 @item ctl-amd restart
7724 Stop the running @i{Amd} (if any), safely wait for it to terminate, and
7725 then start a new process --- only if the previous one died cleanly.
7728 @xref{Run-time Administration}, for more details.
7730 @c ----------------------------------------------------------------
7731 @node ctl-hlfsd, fix-amd-map, ctl-amd, Assorted Tools
7732 @comment node-name, next, previous, up
7736 A script for controlling @i{Hlfsd}, much the same way @file{ctl-amd}
7737 controls @i{Amd}. Use it as follows:
7740 @item ctl-hlfsd start
7741 Start a new @i{Hlfsd} process.
7742 @item ctl-hlfsd stop
7743 Stop the running @i{Hlfsd}.
7744 @item ctl-hlfsd restart
7745 Stop the running @i{Hlfsd} (if any), wait for 10 seconds for it to
7746 terminate, and then start a new process --- only if the previous one
7750 @xref{Hlfsd}, for more details.
7752 @c ----------------------------------------------------------------
7753 @node fix-amd-map, fixmount, ctl-hlfsd, Assorted Tools
7754 @comment node-name, next, previous, up
7755 @section fix-amd-map
7758 Am-utils changed some of the syntax and default values of some
7759 variables. For example, the default value for @samp{$@{os@}} for
7760 Solaris 2.x (aka SunOS 5.x) systems used to be @samp{sos5}, it is now
7761 more automatically generated from @file{config.guess} and its value is
7764 This script converts older @i{Amd} maps to new ones. Use it as follows:
7767 fix-amd-map < @i{old.map} > @i{new.map}
7770 @c ----------------------------------------------------------------
7771 @node fixmount, fixrmtab, fix-amd-map, Assorted Tools
7772 @comment node-name, next, previous, up
7776 @samp{fixmount} is a variant of @b{showmount}(8) that can delete bogus
7777 mount entries in remote @b{mountd}(8) daemons. This is useful to
7778 cleanup otherwise ever-accumulating ``junk''. Use it for example:
7781 fixmount -r @i{host}
7784 See the online manual page for @samp{fixmount} for more details of its
7787 @c ----------------------------------------------------------------
7788 @node fixrmtab, lostaltmail, fixmount, Assorted Tools
7789 @comment node-name, next, previous, up
7793 A script to invalidate @file{/etc/rmtab} entries for hosts named. Also
7794 restart mountd for changes to take effect. Use it for example:
7797 fixrmtab @i{host1} @i{host2} @i{...}
7800 @c ----------------------------------------------------------------
7801 @node lostaltmail, lostaltmail.conf-sample, fixrmtab, Assorted Tools
7802 @comment node-name, next, previous, up
7803 @section lostaltmail
7806 A script used with @i{Hlfsd} to resend any ``lost'' mail. @i{Hlfsd}
7807 redirects mail which cannot be written into the user's home directory to
7808 an alternate directory. This is useful to continue delivering mail,
7809 even if the user's file system was unavailable, full, or over quota.
7810 But, the mail which gets delivered to the alternate directory needs to
7811 be resent to its respective users. This is what the @samp{lostaltmail}
7820 This script needs a configuration file @samp{lostaltmail.conf} set up
7821 with the right parameters to properly work. @xref{Hlfsd}, for more
7824 @c ----------------------------------------------------------------
7825 @node lostaltmail.conf-sample, mk-amd-map, lostaltmail, Assorted Tools
7826 @comment node-name, next, previous, up
7827 @section lostaltmail.conf-sample
7828 @pindex lostaltmail.conf-sample
7829 @cindex lostaltmail; configuration file
7831 This is a text file with configuration parameters needed for the
7832 @samp{lostaltmail} script. The script includes comments explaining each
7833 of the configuration variables. See it for more information. Also
7834 @pxref{Hlfsd} for general information.
7836 @c ----------------------------------------------------------------
7837 @node mk-amd-map, pawd, lostaltmail.conf-sample, Assorted Tools
7838 @comment node-name, next, previous, up
7842 This program converts a normal @i{Amd} map file into an ndbm database
7843 with the same prefix as the named file. Use it as follows:
7846 mk-amd-map @i{mapname}
7849 @c ----------------------------------------------------------------
7850 @node pawd, redhat-ctl-amd, mk-amd-map, Assorted Tools
7851 @comment node-name, next, previous, up
7855 @i{Pawd} is used to print the current working directory, adjusted to
7856 reflect proper paths that can be reused to go through the automounter
7857 for the shortest possible path. In particular, the path printed back
7858 does not include any of @i{Amd}'s local mount points. Using them is
7859 unsafe, because @i{Amd} may unmount managed file systems from the mount
7860 points, and thus including them in paths may not always find the files
7863 Without any arguments, @i{Pawd} will print the automounter adjusted
7864 current working directory. With any number of arguments, it will print
7865 the adjusted path of each one of the arguments.
7867 @c ----------------------------------------------------------------
7868 @node redhat-ctl-amd, wait4amd, pawd, Assorted Tools
7869 @comment node-name, next, previous, up
7870 @section redhat-ctl-amd
7871 @pindex redhat-ctl-amd
7873 This script is similar to @i{ctl-amd} (@pxref{ctl-amd}) but is intended
7874 for Red Hat Linux systems. You can safely copy @i{redhat-ctl-amd} onto
7875 @file{/etc/rc.d/init.d/amd}. The script supplied by @i{Am-utils} is
7876 usually better than the one provided by Red Hat, because the Red Hat
7877 script does not correctly kill @i{Amd} processes: it is too quick to
7878 kill the wrong processes, leaving stale or hung mount points behind.
7880 @c ----------------------------------------------------------------
7881 @node wait4amd, wait4amd2die, redhat-ctl-amd, Assorted Tools
7882 @comment node-name, next, previous, up
7886 A script to wait for @i{Amd} to start on a particular host before
7887 performing an arbitrary command. The command is executed repeatedly,
7888 with 1 second intervals in between. You may interrupt the script using
7889 @samp{^C} (or whatever keyboard sequence your terminal's @samp{intr} function
7895 @item wait4amd saturn amq -p -h saturn
7896 When @i{Amd} is up on host @samp{saturn}, get the process ID of that
7898 @item wait4amd pluto rlogin pluto
7899 Remote login to host @samp{pluto} when @i{Amd} is up on that host. It
7900 is generally necessary to wait for @i{Amd} to properly start and
7901 initialize on a remote host before logging in to it, because otherwise
7902 user home directories may not be accessible across the network.
7903 @item wait4amd pluto
7904 A short-hand version of the previous command, since the most useful
7905 reason for this script is to login to a remote host. I use it very
7906 often when testing out new versions of @i{Amd}, and need to reboot hung
7910 @c ----------------------------------------------------------------
7911 @node wait4amd2die, wire-test, wait4amd, Assorted Tools
7912 @comment node-name, next, previous, up
7913 @section wait4amd2die
7914 @pindex wait4amd2die
7916 This script is used internally by @samp{ctl-amd} when used to restart
7917 @i{Amd}. It waits for @i{Amd} to terminate. If it detected that
7918 @i{Amd} terminated cleanly, this script will return an exist status of
7919 zero. Otherwise, it will return a non-zero exit status.
7921 The script tests for @i{Amd}'s existence once every 5 seconds, six
7922 times, for a total of 30 seconds. It will return a zero exist status as
7923 soon as it detects that @i{Amd} dies.
7925 @c ----------------------------------------------------------------
7926 @node wire-test, , wait4amd2die, Assorted Tools
7927 @comment node-name, next, previous, up
7931 A simple program to test if some of the most basic networking functions
7932 in am-util's library @file{libamu} work. It also tests the combination
7933 of NFS protocol and version number that are supported from the current
7934 host, to a remote one.
7936 For example, in this test a machine which only supports NFS Version 2 is
7937 contacting a remote host that can support the same version, but using
7938 both UDP and TCP. If no host name is specified, @samp{wire-test} will
7939 try @file{localhost}.
7943 Network name is "mcl-lab-net.cs.columbia.edu"
7944 Network number is "128.59.13"
7945 Network name is "old-net.cs.columbia.edu"
7946 Network number is "128.59.16"
7947 My IP address is 0x7f000001.
7948 NFS Version and protocol tests to host "moisil"...
7949 testing vers=2, proto="udp" -> found version 2.
7950 testing vers=3, proto="udp" -> failed!
7951 testing vers=2, proto="tcp" -> found version 2.
7952 testing vers=3, proto="tcp" -> failed!
7955 @c ################################################################
7956 @node Examples, Internals, Assorted Tools, Top
7957 @comment node-name, next, previous, up
7961 * User Filesystems::
7962 * Home Directories::
7963 * Architecture Sharing::
7967 * /defaults with selectors::
7968 * /tftpboot in a chroot-ed environment::
7972 @node User Filesystems, Home Directories, Examples, Examples
7973 @comment node-name, next, previous, up
7974 @section User Filesystems
7975 @cindex User filesystems
7976 @cindex Mounting user filesystems
7978 With more than one fileserver, the directories most frequently
7979 cross-mounted are those containing user home directories. A common
7980 convention used at Imperial College is to mount the user disks under
7981 @t{/home/}@i{machine}.
7983 Typically, the @samp{/etc/fstab} file contained a long list of entries
7987 @i{machine}:/home/@i{machine} /home/@i{machine} nfs ...
7990 for each fileserver on the network.
7992 There are numerous problems with this system. The mount list can become
7993 quite large and some of the machines may be down when a system is
7994 booted. When a new fileserver is installed, @samp{/etc/fstab} must be
7995 updated on every machine, the mount directory created and the filesystem
7998 In many environments most people use the same few workstations, but
7999 it is convenient to go to a colleague's machine and access your own
8000 files. When a server goes down, it can cause a process on a client
8001 machine to hang. By minimizing the mounted filesystems to only include
8002 those actively being used, there is less chance that a filesystem will
8003 be mounted when a server goes down.
8005 The following is a short extract from a map taken from a research fileserver
8006 at Imperial College.
8008 Note the entry for @samp{localhost} which is used for users such as
8009 the operator (@samp{opr}) who have a home directory on most machine as
8010 @samp{/home/localhost/opr}.
8013 /defaults opts:=rw,intr,grpid,nosuid
8014 charm host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \
8015 host==$@{key@};type:=ufs;dev:=/dev/xd0g
8020 localhost type:=link;fs:=$@{host@}
8023 # dylan has two user disks so have a
8024 # top directory in which to mount them.
8026 dylan type:=auto;fs:=$@{map@};pref:=$@{key@}/
8028 dylan/dk2 host!=dylan;type:=nfs;rhost:=dylan;rfs:=/home/$@{key@} \
8029 host==dylan;type:=ufs;dev:=/dev/dsk/2s0
8031 dylan/dk5 host!=dylan;type:=nfs;rhost:=dylan;rfs:=/home/$@{key@} \
8032 host==dylan;type:=ufs;dev:=/dev/dsk/5s0
8035 toytown host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \
8036 host==$@{key@};type:=ufs;dev:=/dev/xy1g
8039 zebedee host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@} \
8040 host==$@{key@};type:=ufs;dev:=/dev/dsk/1s0
8042 # Just for access...
8044 gould type:=auto;fs:=$@{map@};pref:=$@{key@}/
8045 gould/staff host!=gould;type:=nfs;rhost:=gould;rfs:=/home/$@{key@}
8047 gummo host!=$@{key@};type:=nfs;rhost:=$@{key@};rfs:=/home/$@{key@}
8051 This map is shared by most of the machines listed so on those
8052 systems any of the user disks is accessible via a consistent name.
8053 @i{Amd} is started with the following command
8059 Note that when mounting a remote filesystem, the @dfn{automounted}
8060 mount point is referenced, so that the filesystem will be mounted if
8061 it is not yet (at the time the remote @samp{mountd} obtains the file handle).
8063 @node Home Directories, Architecture Sharing, User Filesystems, Examples
8064 @comment node-name, next, previous, up
8065 @section Home Directories
8066 @cindex Home directories
8067 @cindex Example of mounting home directories
8068 @cindex Mount home directories
8070 One convention for home directories is to locate them in @samp{/homes}
8071 so user @samp{jsp}'s home directory is @samp{/homes/jsp}. With more
8072 than a single fileserver it is convenient to spread user files across
8073 several machines. All that is required is a mount-map which converts
8074 login names to an automounted directory.
8076 Such a map might be started by the command:
8079 amd /homes amd.homes
8082 where the map @samp{amd.homes} contained the entries:
8085 /defaults type:=link # All the entries are of type:=link
8086 jsp fs:=/home/charm/jsp
8087 njw fs:=/home/dylan/dk5/njw
8089 phjk fs:=/home/toytown/ai/phjk
8090 sjv fs:=/home/ganymede/sjv
8093 Whenever a login name is accessed in @samp{/homes} a symbolic link
8094 appears pointing to the real location of that user's home directory. In
8095 this example, @samp{/homes/jsp} would appear to be a symbolic link
8096 pointing to @samp{/home/charm/jsp}. Of course, @samp{/home} would also
8097 be an automount point.
8099 This system causes an extra level of symbolic links to be used.
8100 Although that turns out to be relatively inexpensive, an alternative is
8101 to directly mount the required filesystems in the @samp{/homes}
8102 map. The required map is simple, but long, and its creation is best automated.
8103 The entry for @samp{jsp} could be:
8106 jsp -sublink:=$@{key@};rfs:=/home/charm \
8107 host==charm;type:=ufs;dev:=/dev/xd0g \
8108 host!=charm;type:=nfs;rhost:=charm
8111 This map can become quite big if it contains a large number of entries.
8112 By combining two other features of @i{Amd} it can be greatly simplified.
8114 First the UFS partitions should be mounted under the control of
8115 @samp{/etc/fstab}, taking care that they are mounted in the same place
8116 that @i{Amd} would have automounted them. In most cases this would be
8117 something like @samp{/a/@dfn{host}/home/@dfn{host}} and
8118 @samp{/etc/fstab} on host @samp{charm} would have a line:@refill
8121 /dev/xy0g /a/charm/home/charm 4.2 rw,nosuid,grpid 1 5
8124 The map can then be changed to:
8127 /defaults type:=nfs;sublink:=$@{key@};opts:=rw,intr,nosuid,grpid
8128 jsp rhost:=charm;rfs:=/home/charm
8129 njw rhost:=dylan;rfs:=/home/dylan/dk5
8131 phjk rhost:=toytown;rfs:=/home/toytown;sublink:=ai/$@{key@}
8132 sjv rhost:=ganymede;rfs:=/home/ganymede
8135 This map operates as usual on a remote machine (@i{ie} @code{$@{host@}}
8136 not equal to @code{$@{rhost@}}). On the machine where the filesystem is
8137 stored (@i{ie} @code{$@{host@}} equal to @code{$@{rhost@}}), @i{Amd}
8138 will construct a local filesystem mount point which corresponds to the
8139 name of the locally mounted UFS partition. If @i{Amd} is started with
8140 the @code{-r} option then instead of attempting an NFS mount, @i{Amd} will
8141 simply inherit the UFS mount (@pxref{Inheritance Filesystem}). If
8142 @code{-r} is not used then a loopback NFS mount will be made. This type of
8143 mount is known to cause a deadlock on many systems.
8145 @node Architecture Sharing, Wildcard Names, Home Directories, Examples
8146 @comment node-name, next, previous, up
8147 @section Architecture Sharing
8148 @cindex Architecture sharing
8149 @cindex Sharing a fileserver between architectures
8150 @cindex Architecture dependent volumes
8152 @c %At the moment some of the research machines have sets of software
8153 @c %mounted in @samp{/vol}. This contains subdirectories for \TeX,
8154 @c %system sources, local sources, prolog libraries and so on.
8155 Often a filesystem will be shared by machines of different architectures.
8156 Separate trees can be maintained for the executable images for each
8157 architecture, but it may be more convenient to have a shared tree,
8158 with distinct subdirectories.
8160 A shared tree might have the following structure on the fileserver (called
8161 @samp{fserver} in the example):
8170 local/tex/bin/hp9000
8174 In this example, the subdirectories of @samp{local/tex/bin} should be
8175 hidden when accessed via the automount point (conventionally @samp{/vol}).
8176 A mount-map for @samp{/vol} to achieve this would look like:
8179 /defaults sublink:=$@{/key@};rhost:=fserver;type:=link
8180 tex type:=auto;fs:=$@{map@};pref:=$@{key@}/
8181 tex/fonts host!=fserver;type:=nfs;rfs:=/vol/tex \
8182 host==fserver;fs:=/usr/local/tex
8183 tex/lib host!=fserver;type:=nfs;rfs:=/vol/tex \
8184 host==fserver;fs:=/usr/local/tex
8185 tex/bin -sublink:=$@{/key@}/$@{arch@} \
8186 host!=fserver;type:=nfs;rfs:=/vol/tex \
8187 host:=fserver;fs:=/usr/local/tex
8190 When @samp{/vol/tex/bin} is referenced, the current machine architecture
8191 is automatically appended to the path by the @code{$@{sublink@}}
8192 variable. This means that users can have @samp{/vol/tex/bin} in their
8193 @samp{PATH} without concern for architecture dependencies.
8195 @node Wildcard Names, rwho servers, Architecture Sharing, Examples
8196 @comment node-name, next, previous, up
8197 @section Wildcard Names & Replicated Servers
8199 By using the wildcard facility, @i{Amd} can @dfn{overlay} an existing
8200 directory with additional entries.
8201 The system files are usually mounted under @samp{/usr}. If instead,
8202 @i{Amd} is mounted on @samp{/usr}, additional
8203 names can be overlayed to augment or replace names in the ``master'' @samp{/usr}.
8204 A map to do this would have the form:
8207 local type:=auto;fs:=local-map
8208 share type:=auto;fs:=share-map
8209 * -type:=nfs;rfs:=/export/exec/$@{arch@};sublink:="$@{key@}" \
8210 rhost:=fserv1 rhost:=fserv2 rhost:=fserv3
8213 Note that the assignment to @code{$@{sublink@}} is surrounded by double
8214 quotes to prevent the incoming key from causing the map to be
8215 misinterpreted. This map has the effect of directing any access to
8216 @samp{/usr/local} or @samp{/usr/share} to another automount point.
8218 In this example, it is assumed that the @samp{/usr} files are replicated
8219 on three fileservers: @samp{fserv1}, @samp{fserv2} and @samp{fserv3}.
8220 For any references other than to @samp{local} and @samp{share} one of
8221 the servers is used and a symbolic link to
8222 @t{$@{autodir@}/$@{rhost@}/export/exec/$@{arch@}/@i{whatever}} is
8223 returned once an appropriate filesystem has been mounted.@refill
8225 @node rwho servers, /vol, Wildcard Names, Examples
8226 @comment node-name, next, previous, up
8227 @section @samp{rwho} servers
8228 @cindex rwho servers
8229 @cindex Architecture specific mounts
8230 @cindex Example of architecture specific mounts
8232 The @samp{/usr/spool/rwho} directory is a good candidate for automounting.
8233 For efficiency reasons it is best to capture the rwho data on a small
8234 number of machines and then mount that information onto a large number
8235 of clients. The data written into the rwho files is byte order dependent
8236 so only servers with the correct byte ordering can be used by a client:
8240 usr/spool/rwho -byte==little;rfs:=/usr/spool/rwho \
8241 rhost:=vaxA rhost:=vaxB \
8242 || -rfs:=/usr/spool/rwho \
8243 rhost:=sun4 rhost:=hp300
8246 @node /vol, /defaults with selectors, rwho servers, Examples
8247 @comment node-name, next, previous, up
8248 @section @samp{/vol}
8250 @cindex Catch-all mount point
8251 @cindex Generic volume name
8253 @samp{/vol} is used as a catch-all for volumes which do not have other
8256 Below is part of the @samp{/vol} map for the domain @samp{doc.ic.ac.uk}.
8257 The @samp{r+d} tree is used for new or experimental software that needs
8258 to be available everywhere without installing it on all the fileservers.
8259 Users wishing to try out the new software then simply include
8260 @samp{/vol/r+d/@{bin,ucb@}} in their path.@refill
8262 The main tree resides on one host @samp{gould.doc.ic.ac.uk}, which has
8263 different @samp{bin}, @samp{etc}, @samp{lib} and @samp{ucb}
8264 sub-directories for each machine architecture. For example,
8265 @samp{/vol/r+d/bin} for a Sun-4 would be stored in the sub-directory
8266 @samp{bin/sun4} of the filesystem @samp{/usr/r+d}. When it was accessed
8267 a symbolic link pointing to @samp{/a/gould/usr/r+d/bin/sun4} would be
8271 /defaults type:=nfs;opts:=rw,grpid,nosuid,intr,soft
8272 wp -opts:=rw,grpid,nosuid;rhost:=charm \
8273 host==charm;type:=link;fs:=/usr/local/wp \
8274 host!=charm;type:=nfs;rfs:=/vol/wp
8277 src -opts:=rw,grpid,nosuid;rhost:=charm \
8278 host==charm;type:=link;fs:=/usr/src \
8279 host!=charm;type:=nfs;rfs:=/vol/src
8281 r+d type:=auto;fs:=$@{map@};pref:=r+d/
8282 # per architecture bin,etc,lib&ucb...
8283 r+d/bin rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@}
8284 r+d/etc rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@}
8285 r+d/include rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}
8286 r+d/lib rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@}
8287 r+d/man rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}
8288 r+d/src rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}
8289 r+d/ucb rhost:=gould.doc.ic.ac.uk;rfs:=/usr/r+d;sublink:=$@{/key@}/$@{arch@}
8291 pictures -opts:=rw,grpid,nosuid;rhost:=thpfs \
8292 host==thpfs;type:=link;fs:=/nbsd/pictures \
8293 host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=pictures
8295 hades -opts:=rw,grpid,nosuid;rhost:=thpfs \
8296 host==thpfs;type:=link;fs:=/nbsd/hades \
8297 host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=hades
8299 bsd -opts:=rw,grpid,nosuid;arch==hp9000;rhost:=thpfs \
8300 host==thpfs;type:=link;fs:=/nbsd/bsd \
8301 host!=thpfs;type:=nfs;rfs:=/nbsd;sublink:=bsd
8304 @node /defaults with selectors, /tftpboot in a chroot-ed environment, /vol, Examples
8305 @comment node-name, next, previous, up
8306 @section @samp{/defaults} with selectors
8307 @cindex /defaults with selectors
8308 @cindex selectors on default
8310 It is sometimes useful to have different defaults for a given map. To
8311 achieve this, the @samp{/defaults} entry must be able to process normal
8312 selectors. This feature is turned on by setting
8313 @samp{selectors_in_defaults = yes} in the @file{amd.conf} file.
8314 @xref{selectors_in_defaults Parameter}.
8316 In this example, I set different default NFS mount options for hosts
8317 which are running over a slower network link. By setting a smaller size
8318 for the NFS read and write buffer sizes, you can greatly improve remote
8319 file service performance.
8323 wire==slip-net;opts:=rw,intr,rsize=1024,wsize=1024,timeo=20,retrans=10 \
8324 wire!=slip-net;opts:=rw,intr
8327 @node /tftpboot in a chroot-ed environment, , /defaults with selectors, Examples
8328 @comment node-name, next, previous, up
8329 @section @samp{/tftpboot} in a chroot-ed environment
8330 @cindex /tftpboot in a chroot-ed environment
8331 @cindex chroot; /tftpboot example
8333 In this complex example, we attempt to run an @i{Amd} process
8334 @emph{inside} a chroot-ed environment. @samp{tftpd} (Trivial FTP) is
8335 used to trivially retrieve files used to boot X-Terminals, Network
8336 Printers, Network routers, diskless workstations, and other such
8337 devices. For security reasons, @samp{tftpd} (and also @samp{ftpd})
8338 processes are run using the @b{chroot}(2) system call. This provides an
8339 environment for these processes, where access to any files outside the
8340 directory where the chroot-ed process runs is denied.
8342 For example, if you start @samp{tftpd} on your system with
8345 chroot /tftpboot /usr/sbin/tftpd
8349 then the @samp{tftpd} process will not be able to access any files
8350 outside @file{/tftpboot}. This ensures that no one can retrieve files
8351 such as @file{/etc/passwd} and run password crackers on it.
8353 Since the TFTP service works by broadcast, it is necessary to have at
8354 least one TFTP server running on each subnet. If you have lots of files
8355 that you need to make available for @samp{tftp}, and many subnets, it
8356 could take significant amounts of disk space on each host serving them.
8358 A solution we implemented at Columbia University was to have every host
8359 run @samp{tftpd}, but have those servers retrieve the boot files from
8360 two replicated servers. Those replicated servers have special
8361 partitions dedicated to the many network boot files.
8363 We start @i{Amd} as follows:
8366 amd /tftpboot/.amd amd.tftpboot
8369 That is, @i{Amd} is serving the directory @file{/tftpboot/.amd}. The
8370 @samp{tftp} server runs inside @file{/tftpboot} and is chroot-ed in that
8371 directory too. The @file{amd.tftpboot} map looks like:
8375 # Amd /tftpboot directory -> host map
8378 /defaults opts:=nosuid,ro,intr,soft;fs:=/tftpboot/import;type:=nfs
8380 tp host==lol;rfs:=/n/lol/import/tftpboot;type:=lofs \
8381 host==ober;rfs:=/n/ober/misc/win/tftpboot;type:=lofs \
8382 rhost:=ober;rfs:=/n/ober/misc/win/tftpboot \
8383 rhost:=lol;rfs:=/n/lol/import/tftpboot
8386 To help understand this example, I list a few of the file entries that
8387 are created inside @file{/tftpboot}:
8391 dr-xr-xr-x 2 root 512 Aug 30 23:11 .amd
8392 drwxrwsr-x 12 root 512 Aug 30 08:00 import
8393 lrwxrwxrwx 1 root 33 Feb 27 1997 adminpr.cfg -> ./.amd/tp/hplj/adminpr.cfg
8394 lrwxrwxrwx 1 root 22 Dec 5 1996 tekxp -> ./.amd/tp/xterms/tekxp
8395 lrwxrwxrwx 1 root 1 Dec 5 1996 tftpboot -> .
8398 Here is an explanation of each of the entries listed above:
8403 This is the @i{Amd} mount point. Note that you do not need to run a
8404 separate @i{Amd} process for the TFTP service. The @b{chroot}(2) system
8405 call only protects against file access, but the same process can still
8406 serve files and directories inside and outside the chroot-ed
8407 environment, because @i{Amd} itself was not run in chroot-ed mode.
8410 This is the mount point where @i{Amd} will mount the directories
8411 containing the boot files. The map is designed so that remote
8412 directories will be NFS mounted (even if they are already mounted
8413 elsewhere), and local directories are loopback mounted (since they are
8414 not accessible outside the chroot-ed @file{/tftpboot} directory).
8418 Two manually created symbolic links to directories @emph{inside} the
8419 @i{Amd}-managed directory. The crossing of the component @file{tp} will
8420 cause @i{Amd} to automount one of the remote replicas. Once crossed,
8421 access to files inside proceeds as usual. The @samp{adminpr.cfg} is a
8422 configuration file for an HP Laser-Jet 4si printer, and the @samp{tekxp}
8423 is a directory for Tektronix X-Terminal boot files.
8426 This innocent looking symlink is important. Usually, when devices boot
8427 via the TFTP service, they perform the @samp{get file} command to
8428 retrieve @var{file}. However, some devices assume that @samp{tftpd}
8429 does not run in a chroot-ed environment, but rather ``unprotected'', and
8430 thus use a full pathname for files to retrieve, as in @samp{get
8431 /tftpboot/file}. This symlink effectively strips out the leading
8436 @c ################################################################
8437 @node Internals, Acknowledgments & Trademarks, Examples, Top
8438 @comment node-name, next, previous, up
8441 Note that there are more error and logging messages possible than are
8442 listed here. Most of them are self-explanatory. Refer to the program
8443 sources for more details on the rest.
8449 @node Log Messages, , Internals, Internals
8450 @comment node-name, next, previous, up
8451 @section Log Messages
8453 In the following sections a brief explanation is given of some of the
8454 log messages made by @i{Amd}. Where the message is in @samp{typewriter}
8455 font, it corresponds exactly to the message produced by @i{Amd}. Words
8456 in @dfn{italic} are replaced by an appropriate string. Variables,
8457 @code{$@{@i{var}@}}, indicate that the value of the appropriate variable is
8460 Log messages are either sent directly to a file,
8461 or logged via the @b{syslog}(3) mechanism. @xref{log_file Parameter}.
8462 In either case, entries in the file are of the form:
8464 @i{date-string} @i{hostname} @t{amd[}@i{pid}@t{]} @i{message}
8472 @node Fatal errors, Info messages, Log Messages, Log Messages
8473 @comment node-name, next, previous, up
8474 @subsection Fatal errors
8476 @i{Amd} attempts to deal with unusual events. Whenever it is not
8477 possible to deal with such an error, @i{Amd} will log an appropriate
8478 message and, if it cannot possibly continue, will either exit or abort.
8479 These messages are selected by @samp{-x fatal} on the command line.
8480 When @b{syslog}(3) is being used, they are logged with level
8481 @samp{LOG_FATAL}. Even if @i{Amd} continues to operate it is likely to
8482 remain in a precarious state and should be restarted at the earliest
8487 @item Attempting to inherit not-a-filesystem
8488 The prototype mount point created during a filesystem restart did not
8489 contain a reference to the restarted filesystem. This error ``should
8492 @item Can't bind to domain "@i{NIS-domain}"
8493 A specific NIS domain was requested on the command line, but no server
8494 for that domain is available on the local net.
8496 @item Can't determine IP address of this host (@i{hostname})
8497 When @i{Amd} starts it determines its own IP address. If this lookup
8498 fails then @i{Amd} cannot continue. The hostname it looks up is that
8499 obtained returned by @b{gethostname}(2) system call.
8501 @item Can't find root file handle for @i{automount point}
8502 @i{Amd} creates its own file handles for the automount points. When it
8503 mounts itself as a server, it must pass these file handles to the local
8504 kernel. If the filehandle is not obtainable the mount point is ignored.
8505 This error ``should never happen''.
8507 @item Must be root to mount filesystems (euid = @i{euid})
8508 To prevent embarrassment, @i{Amd} makes sure it has appropriate system
8509 privileges. This amounts to having an euid of 0. The check is made
8510 after argument processing complete to give non-root users a chance to
8511 access the @code{-v} option.
8513 @item No work to do - quitting
8514 No automount points were given on the command line and so there is no
8518 While attempting to malloc some memory, the memory space available to
8519 @i{Amd} was exhausted. This is an unrecoverable error.
8521 @item Out of memory in realloc
8522 While attempting to realloc some memory, the memory space available to
8523 @i{Amd} was exhausted. This is an unrecoverable error.
8525 @item cannot create rpc/udp service
8526 Either the NFS or AMQ endpoint could not be created.
8528 @item gethostname: @i{description}
8529 The @b{gethostname}(2) system call failed during startup.
8531 @item host name is not set
8532 The @b{gethostname}(2) system call returned a zero length host name.
8533 This can happen if @i{Amd} is started in single user mode just after
8536 @item ifs_match called!
8537 An internal error occurred while restarting a pre-mounted filesystem.
8538 This error ``should never happen''.
8540 @item mount_afs: @i{description}
8541 An error occurred while @i{Amd} was mounting itself.
8543 @item run_rpc failed
8544 Somehow the main NFS server loop failed. This error ``should never
8547 @item unable to free rpc arguments in amqprog_1
8548 The incoming arguments to the AMQ server could not be free'ed.
8550 @item unable to free rpc arguments in nfs_program_1
8551 The incoming arguments to the NFS server could not be free'ed.
8553 @item unable to register (AMQ_PROGRAM, AMQ_VERSION, udp)
8554 The AMQ server could not be registered with the local portmapper or the
8555 internal RPC dispatcher.
8557 @item unable to register (NFS_PROGRAM, NFS_VERSION, 0)
8558 The NFS server could not be registered with the internal RPC dispatcher.
8562 XXX: This section needs to be updated
8564 @node Info messages, , Fatal errors, Log Messages
8565 @comment node-name, next, previous, up
8566 @subsection Info messages
8568 @i{Amd} generates information messages to record state changes. These
8569 messages are selected by @samp{-x info} on the command line. When
8570 @b{syslog}(3) is being used, they are logged with level @samp{LOG_INFO}.
8572 The messages listed below can be generated and are in a format suitable
8573 for simple statistical analysis. @dfn{mount-info} is the string
8574 that is displayed by @dfn{Amq} in its mount information column and
8575 placed in the system mount table.
8579 @item "@t{$@{@i{path}@}}" forcibly timed out
8580 An automount point has been timed out by the @i{Amq} command.
8582 @item "@t{$@{@i{path}@}}" has timed out
8583 No access to the automount point has been made within the timeout
8586 @item Filehandle denied for "$@{@i{rhost}@}:$@{@i{rfs}@}"
8587 The mount daemon refused to return a file handle for the requested filesystem.
8589 @item Filehandle error for "$@{@i{rhost}@}:$@{@i{rfs}@}": @i{description}
8590 The mount daemon gave some other error for the requested filesystem.
8592 @item Finishing with status @i{exit-status}
8593 @i{Amd} is about to exit with the given exit status.
8595 @item Re-synchronizing cache for map @t{$@{@i{map}@}}
8596 The named map has been modified and the internal cache is being re-synchronized.
8598 @item file server @t{$@{@i{rhost}@}} is down - timeout of "@t{$@{@i{path}@}}" ignored
8599 An automount point has timed out, but the corresponding file server is
8600 known to be down. This message is only produced once for each mount
8601 point for which the server is down.
8603 @item file server @t{$@{@i{rhost}@}} type nfs is down
8604 An NFS file server that was previously up is now down.
8606 @item file server @t{$@{@i{rhost}@}} type nfs is up
8607 An NFS file server that was previously down is now up.
8609 @item file server @t{$@{@i{rhost}@}} type nfs starts down
8610 A new NFS file server has been referenced and is known to be down.
8612 @item file server @t{$@{@i{rhost}@}} type nfs starts up
8613 A new NFS file server has been referenced and is known to be up.
8615 @item mount of "@t{$@{@i{path}@}}" on @t{$@{@i{fs}@}} timed out
8616 Attempts to mount a filesystem for the given automount point have failed
8617 to complete within 30 seconds.
8619 @item @i{mount-info} mounted fstype @t{$@{@i{type}@}} on @t{$@{@i{fs}@}}
8620 A new file system has been mounted.
8622 @item @i{mount-info} restarted fstype @t{$@{@i{type}@}} on @t{$@{@i{fs}@}}
8623 @i{Amd} is using a pre-mounted filesystem to satisfy a mount request.
8625 @item @i{mount-info} unmounted fstype @t{$@{@i{type}@}} from @t{$@{@i{fs}@}}
8626 A file system has been unmounted.
8628 @item @i{mount-info} unmounted fstype @t{$@{@i{type}@}} from @t{$@{@i{fs}@}} link @t{$@{@i{fs}@}}/@t{$@{@i{sublink}@}}
8629 A file system of which only a sub-directory was in use has been unmounted.
8631 @item restarting @i{mount-info} on @t{$@{@i{fs}@}}
8632 A pre-mounted file system has been noted.
8636 XXX: This section needs to be updated
8638 @c ################################################################
8639 @node Acknowledgments & Trademarks, Index, Internals, Top
8640 @comment node-name, next, previous, up
8641 @unnumbered Acknowledgments & Trademarks
8643 Many thanks to the Am-Utils Users
8644 mailing list through the months developing am-utils. These members
8645 have contributed to the discussions, ideas, code and documentation,
8646 and subjected their systems to alpha quality code. Special thanks go
8647 to those @uref{http://www.am-utils.org/docs/am-utils/AUTHORS.txt,authors} who have
8648 submitted patches, and especially to the maintainers:
8651 @item @uref{http://www.cs.sunysb.edu/~ezk,Erez Zadok}
8652 @item @email{ionut AT badula.org,Ion Badulescu}
8653 @item @email{ro AT techfak.uni-bielefeld.de,Rainer Orth}
8654 @item @email{nick.williams AT morganstanley.com,Nick Williams}
8657 Thanks to the Formal Methods Group at Imperial College for suffering
8658 patiently while @i{Amd} was being developed on their machines.
8660 Thanks to the many people who have helped with the development of
8661 @i{Amd}, especially Piete Brooks at the Cambridge University Computing
8662 Lab for many hours of testing, experimentation and discussion.
8664 Thanks to the older @email{amd-workers AT majordomo.glue.umd.edu,Amd
8665 Workers} mailing list (now defunct) members for many suggestions and
8666 bug reports to @i{Amd}.
8670 @b{DEC}, @b{VAX} and @b{Ultrix} are registered trademarks of Digital
8671 Equipment Corporation.
8673 @b{AIX} and @b{IBM} are registered trademarks of International Business
8674 Machines Corporation.
8676 @b{Sun}, @b{NFS} and @b{SunOS} are registered trademarks of Sun
8679 @b{UNIX} is a registered trademark in the USA and other countries,
8680 exclusively licensed through X/Open Company, Ltd.
8682 All other registered trademarks are owned by their respective owners.
8685 @c ################################################################
8686 @node Index, , Acknowledgments & Trademarks, Top
8687 @comment node-name, next, previous, up
8695 @c ====================================================================
8696 @c ISPELL LOCAL WORDS:
8697 @c LocalWords: setfilename amdref overfullrule settitle titlepage titlefont nz
8698 @c LocalWords: authorfont vskip ifinfo iftex cindex unnumberedsec dfn xref vol
8699 @c LocalWords: locationN pxref jpo nott concentrix Sjoerd sjoerd cwi Eitan vuw
8700 @c LocalWords: Mizrotsky eitan shumuji dgux fpx scp hcx metcalf masala hlh OTS
8701 @c LocalWords: Presnell srp cgl Trost trost ogi pyrOSx OSx tubsibr riscix iX
8702 @c LocalWords: Piete pb Lindblad cjl ai umax utek xinu Mitchum D'Souza dsouza
8703 @c LocalWords: mrc apu alliant aviion AViiON fps macII multimax tahoe vax emph
8704 @c LocalWords: mapdefault valA valB valC YPTSDIR ETCDIR substr MAKEDBM YPDBDIR
8705 @c LocalWords: NOPUSH njw dylan dk dylan njw anydir domN achilles mjh pref sel
8706 @c LocalWords: gdef loc loc loc ldots autodir remopts rwho rwho styx styx yoyo
8707 @c LocalWords: noindent gould rvdmount rvdunmount fserver mtmp unioned logfile
8708 @c LocalWords: dmn esac phjk toytown toytown toytown toytown phjk RdDir RdLnk
8709 @c LocalWords: volname attrs netif dougal inaddr hwaddr ec mountmaps passno xy
8710 @c LocalWords: freq dumpset hfs brian florence localinfo fstabs automaps defn
8711 @c LocalWords: localname fsck'd opr gummo sjv ganymede sjv fserv fserv fserv
8712 @c LocalWords: vaxA vaxB wp thpfs nbsd asis ifs amqprog free'ed printindex gov
8713 @c LocalWords: LocalWords syncodeindex Distrib bsdnet lanl AutoMounter acis ic
8714 @c LocalWords: ac uk aix bsd Mullender nl il DG lcs hpux irix ucsf NeXT cse cl
8715 @c LocalWords: mt FX hp ibm mips utils def def Domainname eg hostd getwd tmp
8716 @c LocalWords: subsubsection rw grpid intr noconn nocto nodevs nosuid retrans
8717 @c LocalWords: rsize tcp timeo nounmount utimeout DDEBUG nodaemon fd hostnames
8718 @c LocalWords: pid Amd's pendry vangogh nfsx backoff stats nomap nostats CRIT
8719 @c LocalWords: noinfo clustername RVD dsk dsk amq hostports osver statfs str
8720 @c LocalWords: ou counter's amdmaps proj src tftpboot sh mv cd sbin ypcat inet
8721 @c LocalWords: Getattr getattr localhost fhandles netmask fstype noquota addr
8722 @c LocalWords: exportfs Dumpsets dumpsets pindex ldif fixmount fixrmtab euid
8723 @c LocalWords: lostaltmail realloc netnumber itemx primnetnum primnetname ARG
8724 @c LocalWords: subsnetname subsnetnum netgrp netgroup multitable Shlib dec osf
8725 @c LocalWords: hppa pc bsdi freebsd netbsd openbsd ncr sysv rs acdirmax fsid
8726 @c LocalWords: acdirmin acregmax acregmin actimeo dumbtimr nfsv noac noauto sd
8727 @c LocalWords: nocache nodev noint nosub pgthresh posix rdonly suid symttl mfs
8728 @c LocalWords: AMFS umapfs myftpdir unionfs es mapname mapfile mapfile slocal
8729 @c LocalWords: mailspool saturn saturn notknown lol ober dr xr xr drwxrwsr cfg
8730 @c LocalWords: lrwxrwxrwx adminpr hplj adminpr cfg tekxp xterms tekxp Dupuy tp
8731 @c LocalWords: linkname hlfsddump dirname rmtab pluto rlogin direntry pg vr dn
8732 @c LocalWords: maxmem hlfsdir xmailbox showmount cn amdmap amdmapName resvport
8733 @c LocalWords: objectClass amdmapKey amdmapValue ln powerpc amdmapTimestamp ez
8734 @c LocalWords: moisil FSinfo Libtool Unmounting sublink fileservers NullProc
8735 @c LocalWords: gethostname mount's unmounts linkx remounts unmounting UAs SA's
8736 @c LocalWords: mountpoint mountpoints unescaped UIDs util's overlayed uref EFS
8737 @c LocalWords: serv maxgroups nfsl cachedir copt cfsadmin efs addopts fg ROMs
8738 @c LocalWords: nointr extatt setchapternewpage columnfractions alphaev gnulibc
8739 @c LocalWords: freebsdelf gnuoldld ifhtml defperm nodefperm norrip RRIP rrip
8740 @c LocalWords: noversion attr XXXXXX netgrpd rh mkstemp uid gid noexec mntfs
8741 @c LocalWords: nomnttab optionstr hrtime xdrtrace getpwd proplist redhat ctl
8742 @c LocalWords: texinfo texi ib sp cartouche ified xlatecookie dircategory sc
8743 @c LocalWords: AddInfo suse Novell softlookup ENOENT USB fullybrowsable LDAPv
8744 @c LocalWords: amy ie xfffffe zebedee andrew diskfull hdmail searchable si
8745 @c LocalWords: Orth ESTALE