jail -- imprison process and its descendants
jail [-i] [-u username] path hostname ip-number command ...
The jail utility imprisons a process and all future descendants.
The options are as follows:
-i Output the jail identifier of the newly created jail.
-u username The user name as whom the command should run.
path Directory which is to be the root of the prison.
hostname Hostname of the prison.
ip-number IP number assigned to the prison.
command Pathname of the program which is to be executed.
Jails are typically set up using one of two philosophies: either to constrain
a specific application (possibly running with privilege), or to
create a "virtual system image" running a variety of daemons and services.
In both cases, a fairly complete file system install of FreeBSD
is required, so as to provide the necessary command line tools, daemons,
libraries, application configuration files, etc are available. However,
for a virtual server configuration, a fair amount of additional work is
required so as to configure the "boot" process. This man page documents
the configuration steps necessary to support either of these steps,
althoguh the configuration steps may be refined based on local requirements.
Please see the jail(2) man page for further details.
Setting up a Jail Directory Tree
This example shows how to setup a jail directory tree containing an
entire FreeBSD distribution:
D=/here/is/the/jail
cd /usr/src
mkdir -p $D
make world DESTDIR=$D
cd etc
make distribution DESTDIR=$D
mount_devfs devfs $D/dev
cd $D
ln -sf dev/null kernel
NOTE: It is important that only appropriate device nodes in devfs be
exposed to a jail; access to disk devices in the jail may permit processes
in the jail to bypass the jail sandboxing by modifying files outside
of the jail. See devfs(8) for information on how to use devfs rules
to limit access to entries in the per-jail devfs.
In many cases this example would put far more stuff in the jail than is
needed. In the other extreme case a jail might contain only one single
file: the executable to be run in the jail.
We recommend experimentation and caution that it is a lot easier to start
with a ``fat'' jail and remove things until it stops working, than it is
to start with a ``thin'' jail and add things until it works.
Setting Up a Jail [Toc] [Back]
Do what was described in Setting Up a Jail Directory Tree to build the
jail directory tree. For the sake of this example, we will assume you
built it in /data/jail/192.168.11.100, named for the jailed IP address.
Substitute below as needed with your own directory, IP address, and hostname.
Setting up the Host Environment [Toc] [Back]
First, you will want to set up your real system's environment to be
``jail-friendly''. For consistency, we will refer to the parent box as
the ``host environment'', and to the jailed virtual machine as the ``jail
environment''. Because jail is implemented using IP aliases, one of the
first things to do is to disable IP services on the host system that listen
on all local IP addresses for a service. If a network service is
present in the host environment that binds all available IP addresses
rather than specific IP addresses, it may service requests sent to jail
IP addresses. This means changing inetd(8) to only listen on the appropriate
IP address, and so forth. Add the following to /etc/rc.conf in
the host environment:
sendmail_enable="NO"
inetd_flags="-wW -a 192.168.11.23"
rpcbind_enable="NO"
192.168.11.23 is the native IP address for the host system, in this example.
Daemons that run out of inetd(8) can be easily set to use only the
specified host IP address. Other daemons will need to be manually configured--for
some this is possible through the rc.conf(5) flags entries,
for others it is necessary to modify per-application configuration files,
or to recompile the application. The following frequently deployed services
must have their individual configuration files modified to limit
the application to listening to a specific IP address:
To configure sshd(8), it is necessary to modify /etc/ssh/sshd_config.
To configure sendmail(8), it is necessary to modify
/etc/mail/sendmail.cf.
For named(8), it is necessary to modify /etc/namedb/named.conf.
In addition, a number of services must be recompiled in order to run them
in the host environment. This includes most applications providing services
using rpc(3), such as rpcbind(8,) nfsd(8), and mountd(8). In general,
applications for which it is not possible to specify which IP
address to bind should not be run in the host environment unless they
should also service requests sent to jail IP addresses. Attempting to
serve NFS from the host environment may also cause confusion, and cannot
be easily reconfigured to use only specific IPs, as some NFS services are
hosted directly from the kernel. Any third party network software running
in the host environment should also be checked and configured so
that it does not bind all IP addresses, which would result in those services
also appearing to be offered by the jail environments.
Once these daemons have been disabled or fixed in the host environment,
it is best to reboot so that all daemons are in a known state, to reduce
the potential for confusion later (such as finding that when you send
mail to a jail, and its sendmail is down, the mail is delivered to the
host, etc.)
Configuring the Jail [Toc] [Back]
Start any jails for the first time without configuring the network interface
so that you can clean it up a little and set up accounts. As with
any machine (virtual or not) you will need to set a root password, time
zone, etc. Some of these steps apply only if you intend to run a full
virtual server inside the jail; others apply for both constraining a particular
application or for a virtual server.
Start a shell in the jail:
jail /data/jail/192.168.11.100 testhostname 192.168.11.100 /bin/sh
You will end up with a shell prompt, assuming no errors, within the jail.
You can now run /usr/sbin/sysinstall and do the post-install configuration
to set various configuration options, or perform these actions manually
by editing /etc/rc.conf, etc.
+o Create an empty /etc/fstab to quell startup warnings about
missing fstab (virtual server only)
+o Disable the port mapper (/etc/rc.conf: rpcbind_enable="NO")
(virtual server only)
+o Run newaliases(1) to quell sendmail(8) warnings.
+o Disable interface configuration to quell startup warnings about
ifconfig(8) (network_interfaces="") (virtual server only)
+o Configure /etc/resolv.conf so that name resolution within the
jail will work correctly
+o Set a root password, probably different from the real host system
+o Set the timezone
+o Add accounts for users in the jail environment
+o Install any packages the environment requires
You may also want to perform any package-specific configuration (web
servers, SSH servers, etc), patch up /etc/syslog.conf so it logs as you
would like, etc. If you are not using a virtual server, you may wish to
modify syslogd(8) in the host environment to listen on the syslog socket
in the jail environment; in this example, the syslog socket would be
stored in /data/jail/192.168.11.100/var/run/log.
Exit from the shell, and the jail will be shut down.
Starting the Jail [Toc] [Back]
You are now ready to restart the jail and bring up the environment with
all of its daemons and other programs. If you are running a single
application in the jail, substitute the command used to start the application
for /etc/rc in the examples below. To start a virtual server
environment, /etc/rc is run to launch various daemons and services. To
do this, first bring up the virtual host interface, and then start the
jail's /etc/rc script from within the jail.
NOTE: If you plan to allow untrusted users to have root access inside the
jail, you may wish to consider setting the
security.jail.set_hostname_allowed sysctl variable to 0. Please see the
management discussion later in this document as to why this is a good
idea. If you do decide to set this variable, it must be set before
starting any jails, and once each boot.
ifconfig ed0 inet alias 192.168.11.100/32
mount -t procfs proc /data/jail/192.168.11.100/proc
jail /data/jail/192.168.11.100 testhostname 192.168.11.100 \
/bin/sh /etc/rc
A few warnings will be produced, because most sysctl(8) configuration
variables cannot be set from within the jail, as they are global across
all jails and the host environment. However, it should all work properly.
You should be able to see inetd(8), syslogd(8), and other processes
running within the jail using ps(1), with the `J' flag appearing
beside jailed processes. To see an active list of jails, use the jls(8)
utility. You should also be able to telnet(1) to the hostname or IP
address of the jailed environment, and log in using the accounts you created
previously.
Managing the Jail [Toc] [Back]
Normal machine shutdown commands, such as halt(8), reboot(8), and
shutdown(8), cannot be used successfully within the jail. To kill all
processes in a jail, you may log into the jail and, as root, use one of
the following commands, depending on what you want to accomplish:
kill -TERM -1
kill -KILL -1
This will send the SIGTERM or SIGKILL signals to all processes in the
jail from within the jail. Depending on the intended use of the jail,
you may also want to run /etc/rc.shutdown from within the jail. To kill
processes from outside the jail, use the jexec(8) utility in conjuction
with the one of the kill(1) commands above, or use the killall(1) utility
with the -j option.
The /proc/pid/status file contains, as its last field, the hostname of
the jail in which the process runs, or ``-'' to indicate that the process
is not running within a jail. The ps(1) command also shows a `J' flag
for processes in a jail. However, the hostname for a jail may be, by
default, modified from within the jail, so the /proc status entry is
unreliable by default. To disable the setting of the hostname from
within a jail, set the security.jail.set_hostname_allowed sysctl variable
in the host environment to 0, which will affect all jails. You can have
this sysctl set on each boot using sysctl.conf(5). Just add the following
line to /etc/sysctl.conf:
security.jail.set_hostname_allowed=0
Sysctl MIB Entries [Toc] [Back]
Certain aspects of the jail containments environment may be modified from
the host environment using sysctl(8) MIB variables. Currently, these
variables affect all jails on the system, although in the future this
functionality may be finer grained.
security.jail.set_hostname_allowed
This MIB entry determines whether or not processes within a jail are
allowed to change their hostname via hostname(1) or sethostname(3).
In the current jail implementation, the ability to set the hostname
from within the jail can impact management tools relying on the
accuracy of jail information in /proc. As such, this should be disabled
in environments where privileged access to jails is given out
to untrusted parties.
security.jail.socket_unixiproute_only
The jail functionality binds an IPv4 address to each jail, and limits
access to other network addresses in the IPv4 space that may be
available in the host environment. However, jail is not currently
able to limit access to other network protocol stacks that have not
had jail functionality added to them. As such, by default, processes
within jails may only access protocols in the following
domains: PF_LOCAL, PF_INET, and PF_ROUTE, permitting them access to
UNIX domain sockets, IPv4 addresses, and routing sockets. To enable
access to other domains, this MIB variable may be set to 0.
security.jail.sysvipc_allowed
This MIB entry determines whether or not processes within a jail
have access to System V IPC primitives. In the current jail implementation,
System V primitives share a single namespace across the
host and jail environments, meaning that processes within a jail
would be able to communicate with (and potentially interfere with)
processes outside of the jail, and in other jails. As such, this
functionality is disabled by default, but can be enabled by setting
this MIB entry to 1.
There are currently two MIB related variables that have per-jail settings.
Changes to these variables by a jailed process do not effect the
host environment, only the jail environment. The variables are
kern.securelevel and kern.hostname.
killall(1), newaliases(1), ps(1), chroot(2), jail(2), jail_attach(2),
procfs(5), rc.conf(5), sysctl.conf(5), devfs(8), halt(8), inetd(8),
jexec(8), jls(8), mount_devfs(8), named(8), reboot(8), rpcbind(8),
sendmail(8), shutdown(8), sysctl(8), syslogd(8)
The jail utility appeared in FreeBSD 4.0.
The jail feature was written by Poul-Henning Kamp for R&D Associates
http://www.rndassociates.com/ who contributed it to FreeBSD.
Robert Watson wrote the extended documentation, found a few bugs, added a
few new features, and cleaned up the userland jail environment.
Jail currently lacks the ability to allow access to specific jail information
via ps(1) as opposed to procfs(5). Similarly, it might be a good
idea to add an address alias flag such that daemons listening on all IPs
(INADDR_ANY) will not bind on that address, which would facilitate building
a safe host environment such that host daemons do not impose on services
offered from within jails. Currently, the simplist answer is to
minimize services offered on the host, possibly limiting it to services
offered from inetd(8) which is easily configurable.
FreeBSD 5.2.1 April 8, 2003 FreeBSD 5.2.1 [ Back ] |
