2 .\" Copyright (c) 2000, 2003 Robert N. M. Watson
3 .\" Copyright (c) 2008 James Gritton
4 .\" All rights reserved.
6 .\" Redistribution and use in source and binary forms, with or without
7 .\" modification, are permitted provided that the following conditions
9 .\" 1. Redistributions of source code must retain the above copyright
10 .\" notice, this list of conditions and the following disclaimer.
11 .\" 2. Redistributions in binary form must reproduce the above copyright
12 .\" notice, this list of conditions and the following disclaimer in the
13 .\" documentation and/or other materials provided with the distribution.
15 .\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 .\" ----------------------------------------------------------------------------
29 .\" "THE BEER-WARE LICENSE" (Revision 42):
30 .\" <phk@FreeBSD.org> wrote this file. As long as you retain this notice you
31 .\" can do whatever you want with this stuff. If we meet some day, and you think
32 .\" this stuff is worth it, you can buy me a beer in return. Poul-Henning Kamp
33 .\" ----------------------------------------------------------------------------
42 .Nd "create or modify a system jail"
47 .Op Fl l u Ar username | Fl U Ar username
49 .Op Ar parameter=value ...
54 .Op Fl s Ar securelevel
55 .Op Fl l u Ar username | Fl U Ar username
56 .Op Ar path hostname [ip[,..]] command ...
62 utility creates a new jail or modifies an existing jail, optionally
63 imprisoning the current process (and future descendants) inside it.
65 The options are as follows:
66 .Bl -tag -width indent
68 Allow making changes to a dying jail.
74 and add all IP addresses returned by the resolver
77 addresses for this prison.
78 This may affect default address selection for outgoing IPv4 connections
80 The address first returned by the resolver for each address family
81 will be used as primary address.
86 parameters further down for details.
88 Output the jail identifier of the newly created jail.
91 This is deprecated and is equivalent to setting the
97 file, containing jail identifier, path, hostname, IP and
98 command used to start the jail.
100 Run program in the clean environment.
101 The environment is discarded except for
102 .Ev HOME , SHELL , TERM
108 are set to the target login's default values.
110 is set to the target login.
112 is imported from the current environment.
113 The environment variables from the login class capability database for the
114 target login are also set.
115 .It Fl s Ar securelevel
118 MIB entry to the specified value inside the newly created jail.
119 This is deprecated and is equivalent to setting the
123 The user name from host environment as whom the
127 The user name from jailed environment as whom the
136 parameters (if specified) must not refer to an existing jail.
138 Modify an existing jail.
143 parameters must exist and refer to an existing jail.
145 Create a jail if it does not exist, or modify a jail if it does exist.
149 specified by jid or name.
150 All jailed processes are killed, and all children of this jail are also
159 options must be specified.
164 form, following the options.
165 Some parameters are boolean, and do not have a value but are set by the
166 name alone with or without a
172 Any parameters not set will be given default values, often based on the
177 specifies that the current process should enter the new (or modified) jail,
178 and run the specified command.
179 It must be the last parameter specified, because it includes not only
180 the value following the
182 sign, but also passes the rest of the arguments to the command.
184 Instead of supplying named
186 four fixed parameters may be supplied in order on the command line:
196 parameters aren't in this list, this mode will always create a new jail, and
201 options don't apply (and must not exist).
203 Jails have a set a core parameters, and modules can add their own jail
205 The current set of available parameters can be retrieved via
206 .Dq Nm sysctl Fl d Va security.jail.param .
207 The core parameters are:
208 .Bl -tag -width indent
211 This will be assigned automatically to a new jail (or can be explicitly
212 set), and can be used to identify the jail for later modification, or
219 This is an arbitrary string that identifies a jail (except it may not
224 it can be passed to later
232 is supplied, a default is assumed that is the same as the
235 Directory which is to be the root of the prison.
238 (if any) is run from this directory, as are commands from
241 A comma-separated list of IPv4 addresses assigned to the prison.
242 If this is set, the jail is restricted to using only these addresses.
243 Any attempts to use other addresses fail, and attempts to use wildcard
244 addresses silently use the jailed address instead.
245 For IPv4 the first address given will be kept used as the source address
246 in case source address selection on unbound sockets cannot find a better
248 It is only possible to start multiple jails with the same IP address,
249 if none of the jails has more than this single overlapping IP address
252 A boolean option to change the formerly mentioned behaviour and disable
253 IPv4 source address selection for the prison in favour of the primary
254 IPv4 address of the jail.
255 Source address selection is enabled by default for all jails and a
257 setting of a parent jail is not inherited for any child jails.
259 Control the availability of IPv4 addresses.
262 to allow unrestricted access to all system addresses,
264 to restrict addresses via
268 to stop the jail from using IPv4 entirely.
271 parameter implies a value of
273 .It Va ip6.addr , Va ip6.saddrsel , Va ip6
274 A set of IPv6 options for the prison, the counterparts to
281 Hostname of the prison.
282 Other similar parameters are
283 .Va host.domainname ,
288 Set the origin of hostname and related information.
291 to use the system information and
293 for the jail to use the information from the above fields.
294 Setting any of the above fields implies a value of
297 The value of the jail's
300 A jail never has a lower securelevel than the default system, but by
301 setting this parameter it may have a higher one.
302 If the system securelevel is changed, any jail securelevels will be at
305 The number of the devfs ruleset that is enforced for mounting devfs in
307 A value of zero (default) means no ruleset is enforced.
308 Descendant jails inherit the parent jail's devfs ruleset enforcement.
309 Mounting devfs inside a jail is possible only if the
312 .Va allow.mount.devfs
313 permissions are effective and
315 is set to a value lower than 2.
316 Devfs rules and rulesets cannot be viewed or modified from inside a jail.
318 The number of child jails allowed to be created by this jail (or by
319 other jails under this jail).
320 This limit is zero by default, indicating the jail is not allowed to
323 .Va "Hierarchical Jails"
324 section for more information.
326 The number of descendents of this jail, including its own child jails
327 and any jails created under them.
328 .It Va enforce_statfs
329 This determines which information processes in a jail are able to get
331 It affects the behaviour of the following syscalls:
337 (as well as similar compatibility syscalls).
338 When set to 0, all mount points are available without any restrictions.
339 When set to 1, only mount points below the jail's chroot directory are
341 In addition to that, the path to the jail's chroot directory is removed
342 from the front of their pathnames.
343 When set to 2 (default), above syscalls can operate only on a mount-point
344 where the jail's chroot directory is located.
346 Setting this boolean parameter allows a jail to exist without any
348 Normally, a jail is destroyed as its last process exits.
349 A new jail must have either the
353 pseudo-parameter set.
355 The ID of the cpuset associated with this jail (read-only).
357 This is true if the jail is in the process of shutting down (read-only).
361 of the parent of this jail, or zero if this is a top-level jail
364 Some restrictions of the jail environment may be set on a per-jail
366 With the exception of
367 .Va allow.set_hostname ,
368 these boolean parameters are off by default.
369 .Bl -tag -width indent
370 .It Va allow.set_hostname
371 The jail's hostname may be changed via
376 A process within the jail has access to System V IPC primitives.
377 In the current jail implementation, System V primitives share a single
378 namespace across the host and jail environments, meaning that processes
379 within a jail would be able to communicate with (and potentially interfere
380 with) processes outside of the jail, and in other jails.
381 .It Va allow.raw_sockets
382 The prison root is allowed to create raw sockets.
383 Setting this parameter allows utilities like
387 to operate inside the prison.
388 If this is set, the source IP addresses are enforced to comply
389 with the IP address bound to the jail, regardless of whether or not
392 flag has been set on the socket.
393 Since raw sockets can be used to configure and interact with various
394 network subsystems, extra caution should be used where privileged access
395 to jails is given out to untrusted parties.
397 Normally, privileged users inside a jail are treated as unprivileged by
399 When this parameter is set, such users are treated as privileged, and
400 may manipulate system file flags subject to the usual constraints on
401 .Va kern.securelevel .
403 privileged users inside the jail will be able to mount and unmount file
404 system types marked as jail-friendly.
407 command can be used to find file system types available for mount from
409 This permission is effective only if
411 is set to a value lower than 2.
412 .It Va allow.mount.devfs
413 privileged users inside the jail will be able to mount and unmount the
415 This permission is effective only together with
419 is set to a value lower than 2.
420 Please consider restricting the devfs ruleset with the
423 .It Va allow.mount.nullfs
424 privileged users inside the jail will be able to mount and unmount the
426 This permission is effective only together with
430 is set to a value lower than 2.
431 .It Va allow.mount.procfs
432 privileged users inside the jail will be able to mount and unmount the
434 This permission is effective only together with
438 is set to a value lower than 2.
439 .It Va allow.mount.zfs
440 privileged users inside the jail will be able to mount and unmount the
442 This permission is effective only together with
446 is set to a value lower than 2.
449 for information on how to configure the ZFS filesystem to operate from
452 The prison root may administer quotas on the jail's filesystem(s).
453 This includes filesystems that the jail may share with other jails or
454 with non-jailed parts of the system.
455 .It Va allow.socket_af
456 Sockets within a jail are normally restricted to IPv4, IPv6, local
457 (UNIX), and route. This allows access to other protocol stacks that
458 have not had jail functionality added to them.
462 Jails are typically set up using one of two philosophies: either to
463 constrain a specific application (possibly running with privilege), or
465 .Dq "virtual system image"
466 running a variety of daemons and services.
467 In both cases, a fairly complete file system install of
470 required, so as to provide the necessary command line tools, daemons,
471 libraries, application configuration files, etc.
472 However, for a virtual server configuration, a fair amount of
473 additional work is required so as to configure the
476 This manual page documents the configuration steps necessary to support
477 either of these steps, although the configuration steps may be
478 refined based on local requirements.
480 .Ss "Setting up a Jail Directory Tree"
481 To set up a jail directory tree containing an entire
483 distribution, the following
485 command script can be used:
490 make world DESTDIR=$D
491 make distribution DESTDIR=$D
492 mount -t devfs devfs $D/dev
495 NOTE: It is important that only appropriate device nodes in devfs be
496 exposed to a jail; access to disk devices in the jail may permit processes
497 in the jail to bypass the jail sandboxing by modifying files outside of
501 for information on how to use devfs rules to limit access to entries
502 in the per-jail devfs.
503 A simple devfs ruleset for jails is available as ruleset #4 in
504 .Pa /etc/defaults/devfs.rules .
506 In many cases this example would put far more in the jail than needed.
507 In the other extreme case a jail might contain only one file:
508 the executable to be run in the jail.
510 We recommend experimentation and caution that it is a lot easier to
513 jail and remove things until it stops working,
514 than it is to start with a
516 jail and add things until it works.
517 .Ss "Setting Up a Jail"
518 Do what was described in
519 .Sx "Setting Up a Jail Directory Tree"
520 to build the jail directory tree.
521 For the sake of this example, we will
522 assume you built it in
523 .Pa /data/jail/192.0.2.100 ,
524 named for the jailed IP address.
525 Substitute below as needed with your
526 own directory, IP address, and hostname.
527 .Ss "Setting up the Host Environment"
528 First, you will want to set up your real system's environment to be
530 For consistency, we will refer to the parent box as the
531 .Dq "host environment" ,
532 and to the jailed virtual machine as the
533 .Dq "jail environment" .
534 Since jail is implemented using IP aliases, one of the first things to do
535 is to disable IP services on the host system that listen on all local
536 IP addresses for a service.
537 If a network service is present in the host environment that binds all
538 available IP addresses rather than specific IP addresses, it may service
539 requests sent to jail IP addresses if the jail did not bind the port.
542 to only listen on the
543 appropriate IP address, and so forth.
546 in the host environment:
547 .Bd -literal -offset indent
549 inetd_flags="-wW -a 192.0.2.23"
554 is the native IP address for the host system, in this example.
555 Daemons that run out of
557 can be easily set to use only the specified host IP address.
559 will need to be manually configured\(emfor some this is possible through
562 flags entries; for others it is necessary to modify per-application
563 configuration files, or to recompile the applications.
564 The following frequently deployed services must have their individual
565 configuration files modified to limit the application to listening
566 to a specific IP address:
570 it is necessary to modify
571 .Pa /etc/ssh/sshd_config .
575 it is necessary to modify
576 .Pa /etc/mail/sendmail.cf .
580 it is necessary to modify
581 .Pa /etc/namedb/named.conf .
583 In addition, a number of services must be recompiled in order to run
584 them in the host environment.
585 This includes most applications providing services using
592 In general, applications for which it is not possible to specify which
593 IP address to bind should not be run in the host environment unless they
594 should also service requests sent to jail IP addresses.
596 NFS from the host environment may also cause confusion, and cannot be
597 easily reconfigured to use only specific IPs, as some NFS services are
598 hosted directly from the kernel.
599 Any third-party network software running
600 in the host environment should also be checked and configured so that it
601 does not bind all IP addresses, which would result in those services' also
602 appearing to be offered by the jail environments.
605 these daemons have been disabled or fixed in the host environment, it is
606 best to reboot so that all daemons are in a known state, to reduce the
607 potential for confusion later (such as finding that when you send mail
608 to a jail, and its sendmail is down, the mail is delivered to the host,
610 .Ss "Configuring the Jail"
611 Start any jail for the first time without configuring the network
612 interface so that you can clean it up a little and set up accounts.
614 with any machine (virtual or not) you will need to set a root password, time
616 Some of these steps apply only if you intend to run a full virtual server
617 inside the jail; others apply both for constraining a particular application
618 or for running a virtual server.
620 Start a shell in the jail:
621 .Bd -literal -offset indent
622 jail -c path=/data/jail/192.0.2.100 host.hostname=testhostname \\
623 ip4.addr=192.0.2.100 command=/bin/sh
626 Assuming no errors, you will end up with a shell prompt within the jail.
628 .Pa /usr/sbin/sysinstall
629 and do the post-install configuration to set various configuration options,
630 or perform these actions manually by editing
634 .Bl -bullet -offset indent -compact
638 to quell startup warnings about missing fstab (virtual server only)
640 Disable the port mapper
642 .Li rpcbind_enable="NO" )
643 (virtual server only)
647 so that name resolution within the jail will work correctly
655 Disable interface configuration to quell startup warnings about
657 .Pq Li network_interfaces=""
658 (virtual server only)
660 Set a root password, probably different from the real host system
664 Add accounts for users in the jail environment
666 Install any packages the environment requires
669 You may also want to perform any package-specific configuration (web servers,
670 SSH servers, etc), patch up
672 so it logs as you would like, etc.
673 If you are not using a virtual server, you may wish to modify
675 in the host environment to listen on the syslog socket in the jail
676 environment; in this example, the syslog socket would be stored in
677 .Pa /data/jail/192.0.2.100/var/run/log .
679 Exit from the shell, and the jail will be shut down.
680 .Ss "Starting the Jail"
681 You are now ready to restart the jail and bring up the environment with
682 all of its daemons and other programs.
683 If you are running a single application in the jail, substitute the
684 command used to start the application for
686 in the examples below.
687 To start a virtual server environment,
689 is run to launch various daemons and services.
690 To do this, first bring up the
691 virtual host interface, and then start the jail's
693 script from within the jail.
694 .Bd -literal -offset indent
695 ifconfig ed0 inet alias 192.0.2.100/32
696 mount -t procfs proc /data/jail/192.0.2.100/proc
697 jail -c path=/data/jail/192.0.2.100 host.hostname=testhostname \\
698 ip4.addr=192.0.2.100 command=/bin/sh /etc/rc
701 A few warnings will be produced, because most
703 configuration variables cannot be set from within the jail, as they are
704 global across all jails and the host environment.
705 However, it should all
707 You should be able to see
710 and other processes running within the jail using
714 flag appearing beside jailed processes.
715 To see an active list of jails, use the
718 You should also be able to
720 to the hostname or IP address of the jailed environment, and log
721 in using the accounts you created previously.
723 It is possible to have jails started at boot time.
728 for more information.
731 jail script provides a flexible system to start/stop jails:
735 /etc/rc.d/jail start myjail
736 /etc/rc.d/jail stop myjail
738 .Ss "Managing the Jail"
739 Normal machine shutdown commands, such as
744 cannot be used successfully within the jail.
745 To kill all processes in a
746 jail, you may log into the jail and, as root, use one of the following
747 commands, depending on what you want to accomplish:
748 .Bd -literal -offset indent
757 signals to all processes in the jail from within the jail.
759 the intended use of the jail, you may also want to run
761 from within the jail.
762 To kill processes from outside the jail, use the
764 utility in conjunction with the one of the
767 You may also remove the jail with
770 which will killall the jail's processes with
774 .Pa /proc/ Ns Ar pid Ns Pa /status
775 file contains, as its last field, the name of the jail in which the
778 to indicate that the process is not running within a jail.
783 flag for processes in a jail.
785 You can also list/kill processes based on their jail ID.
786 To show processes and their jail ID, use the following command:
788 .Dl "ps ax -o pid,jid,args"
790 To show and then kill processes in jail number 3 use the following commands:
791 .Bd -literal -offset indent
798 .Ss "Jails and File Systems"
799 It is not possible to
803 any file system inside a jail unless the file system is marked
804 jail-friendly, the jail's
806 parameter is set and the jail's
808 parameter is lower than 2.
810 Multiple jails sharing the same file system can influence each other.
811 For example a user in one jail can fill the file system also
812 leaving no space for processes in the other jail.
815 to prevent this will not work either as the file system quotas
816 are not aware of jails but only look at the user and group IDs.
817 This means the same user ID in two jails share the same file
819 One would need to use one file system per jail to make this work.
820 .Ss "Sysctl MIB Entries"
822 .Va security.jail.jailed
823 can be used to determine if a process is running inside a jail (value
824 is one) or not (value is zero).
827 .Va security.jail.max_af_ips
828 determines how may address per address family a prison may have.
831 Some MIB variables have per-jail settings.
832 Changes to these variables by a jailed process do not effect the host
833 environment, only the jail environment.
835 .Va kern.securelevel ,
837 .Va kern.domainname ,
841 .Ss "Hierarchical Jails"
844 parameter, processes within a jail may be able to create jails of their own.
845 These child jails are kept in a hierarchy, with jails only able to see and/or
846 modify the jails they created (or those jails' children).
847 Each jail has a read-only
849 parameter, containing the
851 of the jail that created it; a
853 of 0 indicates the jail is a child of the current jail (or is a top-level
854 jail if the current process isn't jailed).
856 Jailed processes are not allowed to confer greater permissions than they
857 themselves are given, e.g. if a jail is created with
859 it is not able to create a jail with
862 Similarly, such restrictions as
866 may not be bypassed in child jails.
868 A child jail may in turn create its own child jails if its own
870 parameter is set (remember it is zero by default).
871 These jails are visible to and can be modified by their parent and all
874 Jail names reflect this hierarchy, with a full name being an MIB-type string
876 For example, if a base system process creates a jail
878 and a process under that jail creates another jail
880 then the second jail will be seen as
882 in the base system (though it is only seen as
884 to any processes inside jail
886 Jids on the other hand exist in a single space, and each jail must have a
889 Like the names, a child jail's
891 is relative to its creator's own
893 This is by virtue of the child jail being created in the chrooted
894 environment of the first jail.
928 Hierarchical/extensible jails were introduced in
932 The jail feature was written by
933 .An Poul-Henning Kamp
935 .Pa http://www.rndassociates.com/
936 who contributed it to
940 wrote the extended documentation, found a few bugs, added
941 a few new features, and cleaned up the userland jail environment.
944 added multi-IP jail support for IPv4 and IPv6 based on a patch
946 .An Pawel Jakub Dawidek
950 added the extensible jail parameters and hierarchical jails.
952 Jail currently lacks the ability to allow access to
953 specific jail information via
957 Similarly, it might be a good idea to add an
958 address alias flag such that daemons listening on all IPs
960 will not bind on that address, which would facilitate building a safe
961 host environment such that host daemons do not impose on services offered
963 Currently, the simplest answer is to minimize services
964 offered on the host, possibly limiting it to services offered from
966 which is easily configurable.
968 Great care should be taken when managing directories visible within the jail.
969 For example, if a jailed process has its current working directory set to a
970 directory that is moved out of the jail's chroot, then the process may gain
971 access to the file space outside of the jail.
972 It is recommended that directories always be copied, rather than moved, out