man/man7/namespaces.7

'\" t
.\" Copyright (c) 2013, 2016, 2017 by Michael Kerrisk <mtk.manpages@gmail.com>
.\" and Copyright (c) 2012 by Eric W. Biederman <ebiederm@xmission.com>
.\"
.\" SPDX-License-Identifier: Linux-man-pages-copyleft
.\"
.\"
.TH namespaces 7 (date) "Linux man-pages (unreleased)"
.SH NAME
namespaces \- overview of Linux namespaces
.SH DESCRIPTION
A namespace wraps a global system resource in an abstraction that
makes it appear to the processes within the namespace that they
have their own isolated instance of the global resource.
Changes to the global resource are visible to other processes
that are members of the namespace, but are invisible to other processes.
One use of namespaces is to implement containers.
.P
This page provides pointers to information on the various namespace types,
describes the associated
.I /proc
files, and summarizes the APIs for working with namespaces.
.\"
.SS Namespace types
The following table shows the namespace types available on Linux.
The second column of the table shows the flag value that is used to specify
the namespace type in various APIs.
The third column identifies the manual page that provides details
on the namespace type.
The last column is a summary of the resources that are isolated by
the namespace type.
.P
.TS
lB lB lB      lB
l  lB lw(21n) lx.
Namespace	Flag	Page	Isolates
_
Cgroup	CLONE_NEWCGROUP	\fBcgroup_namespaces\fP(7)	T{
Cgroup root directory
T}
IPC	CLONE_NEWIPC	\fBipc_namespaces\fP(7)	T{
System V IPC,
POSIX message queues
T}
Network	CLONE_NEWNET	\fBnetwork_namespaces\fP(7)	T{
Network devices,
stacks, ports, etc.
T}
Mount	CLONE_NEWNS	\fBmount_namespaces\fP(7)	Mount points
PID	CLONE_NEWPID	\fBpid_namespaces\fP(7)	Process IDs
Time	CLONE_NEWTIME	\fBtime_namespaces\fP(7)	T{
Boot and monotonic
clocks
T}
User	CLONE_NEWUSER	\fBuser_namespaces\fP(7)	T{
User and group IDs
T}
UTS	CLONE_NEWUTS	\fButs_namespaces\fP(7)	T{
Hostname and NIS
domain name
T}
.TE
.\"
.\" ==================== The namespaces API ====================
.\"
.SS The namespaces API
As well as various
.I /proc
files described below,
the namespaces API includes the following system calls:
.TP
.BR clone (2)
The
.BR clone (2)
system call creates a new process.
If the
.I flags
argument of the call specifies one or more of the
.B CLONE_NEW*
flags listed above, then new namespaces are created for each flag,
and the child process is made a member of those namespaces.
(This system call also implements a number of features
unrelated to namespaces.)
.TP
.BR setns (2)
The
.BR setns (2)
system call allows the calling process to join an existing namespace.
The namespace to join is specified via a file descriptor that refers to
one of the
.IR /proc/ pid /ns
files described below.
.TP
.BR unshare (2)
The
.BR unshare (2)
system call moves the calling process to a new namespace.
If the
.I flags
argument of the call specifies one or more of the
.B CLONE_NEW*
flags listed above, then new namespaces are created for each flag,
and the calling process is made a member of those namespaces.
(This system call also implements a number of features
unrelated to namespaces.)
.TP
.BR ioctl (2)
Various
.BR ioctl (2)
operations can be used to discover information about namespaces.
These operations are described in
.BR ioctl_nsfs (2).
.P
Creation of new namespaces using
.BR clone (2)
and
.BR unshare (2)
in most cases requires the
.B CAP_SYS_ADMIN
capability, since, in the new namespace,
the creator will have the power to change global resources
that are visible to other processes that are subsequently created in,
or join the namespace.
User namespaces are the exception: since Linux 3.8,
no privilege is required to create a user namespace.
.\"
.\" ==================== The /proc/[pid]/ns/ directory ====================
.\"
.SS The \fI/proc/\fPpid\fI/ns/\fP directory
Each process has a
.IR /proc/ pid /ns/
.\" See commit 6b4e306aa3dc94a0545eb9279475b1ab6209a31f
subdirectory containing one entry for each namespace that
supports being manipulated by
.BR setns (2):
.P
.in +4n
.EX
$ \fBls \-l /proc/$$/ns | awk \[aq]{print $1, $9, $10, $11}\[aq]\fP
total 0
lrwxrwxrwx. cgroup \-> cgroup:[4026531835]
lrwxrwxrwx. ipc \-> ipc:[4026531839]
lrwxrwxrwx. mnt \-> mnt:[4026531840]
lrwxrwxrwx. net \-> net:[4026531969]
lrwxrwxrwx. pid \-> pid:[4026531836]
lrwxrwxrwx. pid_for_children \-> pid:[4026531834]
lrwxrwxrwx. time \-> time:[4026531834]
lrwxrwxrwx. time_for_children \-> time:[4026531834]
lrwxrwxrwx. user \-> user:[4026531837]
lrwxrwxrwx. uts \-> uts:[4026531838]
.EE
.in
.P
Bind mounting (see
.BR mount (2))
one of the files in this directory
to somewhere else in the filesystem keeps
the corresponding namespace of the process specified by
.I pid
alive even if all processes currently in the namespace terminate.
.P
Opening one of the files in this directory
(or a file that is bind mounted to one of these files)
returns a file handle for
the corresponding namespace of the process specified by
.IR pid .
As long as this file descriptor remains open,
the namespace will remain alive,
even if all processes in the namespace terminate.
The file descriptor can be passed to
.BR setns (2).
.P
In Linux 3.7 and earlier, these files were visible as hard links.
Since Linux 3.8,
.\" commit bf056bfa80596a5d14b26b17276a56a0dcb080e5
they appear as symbolic links.
If two processes are in the same namespace,
then the device IDs and inode numbers of their
.IR /proc/ pid /ns/ xxx
symbolic links will be the same; an application can check this using the
.I stat.st_dev
.\" Eric Biederman: "I reserve the right for st_dev to be significant
.\" when comparing namespaces."
.\" https://lore.kernel.org/lkml/87poky5ca9.fsf@xmission.com/
.\" Re: Documenting the ioctl interfaces to discover relationships...
.\" Date: Mon, 12 Dec 2016 11:30:38 +1300
and
.I stat.st_ino
fields returned by
.BR stat (2).
The content of this symbolic link is a string containing
the namespace type and inode number as in the following example:
.P
.in +4n
.EX
$ \fBreadlink /proc/$$/ns/uts\fP
uts:[4026531838]
.EE
.in
.P
The symbolic links in this subdirectory are as follows:
.TP
.IR /proc/ pid /ns/cgroup " (since Linux 4.6)"
This file is a handle for the cgroup namespace of the process.
.TP
.IR /proc/ pid /ns/ipc " (since Linux 3.0)"
This file is a handle for the IPC namespace of the process.
.TP
.IR /proc/ pid /ns/mnt " (since Linux 3.8)"
.\" commit 8823c079ba7136dc1948d6f6dcb5f8022bde438e
This file is a handle for the mount namespace of the process.
.TP
.IR /proc/ pid /ns/net " (since Linux 3.0)"
This file is a handle for the network namespace of the process.
.TP
.IR /proc/ pid /ns/pid " (since Linux 3.8)"
.\" commit 57e8391d327609cbf12d843259c968b9e5c1838f
This file is a handle for the PID namespace of the process.
This handle is permanent for the lifetime of the process
(i.e., a process's PID namespace membership never changes).
.TP
.IR /proc/ pid /ns/pid_for_children " (since Linux 4.12)"
.\" commit eaa0d190bfe1ed891b814a52712dcd852554cb08
This file is a handle for the PID namespace of
child processes created by this process.
This can change as a consequence of calls to
.BR unshare (2)
and
.BR setns (2)
(see
.BR pid_namespaces (7)),
so the file may differ from
.IR /proc/ pid /ns/pid .
The symbolic link gains a value only after the first child process
is created in the namespace.
(Beforehand,
.BR readlink (2)
of the symbolic link will return an empty buffer.)
.TP
.IR /proc/ pid /ns/time " (since Linux 5.6)"
This file is a handle for the time namespace of the process.
.TP
.IR /proc/ pid /ns/time_for_children " (since Linux 5.6)"
This file is a handle for the time namespace of
child processes created by this process.
This can change as a consequence of calls to
.BR unshare (2)
and
.BR setns (2)
(see
.BR time_namespaces (7)),
so the file may differ from
.IR /proc/ pid /ns/time .
.TP
.IR /proc/ pid /ns/user " (since Linux 3.8)"
.\" commit cde1975bc242f3e1072bde623ef378e547b73f91
This file is a handle for the user namespace of the process.
.TP
.IR /proc/ pid /ns/uts " (since Linux 3.0)"
This file is a handle for the UTS namespace of the process.
.P
Permission to dereference or read
.RB ( readlink (2))
these symbolic links is governed by a ptrace access mode
.B PTRACE_MODE_READ_FSCREDS
check; see
.BR ptrace (2).
.\"
.\" ==================== The /proc/sys/user directory ====================
.\"
.SS The \fI/proc/sys/user\fP directory
The files in the
.I /proc/sys/user
directory (which is present since Linux 4.9) expose limits
on the number of namespaces of various types that can be created.
The files are as follows:
.TP
.I max_cgroup_namespaces
The value in this file defines a per-user limit on the number of
cgroup namespaces that may be created in the user namespace.
.TP
.I max_ipc_namespaces
The value in this file defines a per-user limit on the number of
ipc namespaces that may be created in the user namespace.
.TP
.I max_mnt_namespaces
The value in this file defines a per-user limit on the number of
mount namespaces that may be created in the user namespace.
.TP
.I max_net_namespaces
The value in this file defines a per-user limit on the number of
network namespaces that may be created in the user namespace.
.TP
.I max_pid_namespaces
The value in this file defines a per-user limit on the number of
PID namespaces that may be created in the user namespace.
.TP
.IR max_time_namespaces " (since Linux 5.7)"
.\" commit eeec26d5da8248ea4e240b8795bb4364213d3247
The value in this file defines a per-user limit on the number of
time namespaces that may be created in the user namespace.
.TP
.I max_user_namespaces
The value in this file defines a per-user limit on the number of
user namespaces that may be created in the user namespace.
.TP
.I max_uts_namespaces
The value in this file defines a per-user limit on the number of
uts namespaces that may be created in the user namespace.
.P
Note the following details about these files:
.IP \[bu] 3
The values in these files are modifiable by privileged processes.
.IP \[bu]
The values exposed by these files are the limits for the user namespace
in which the opening process resides.
.IP \[bu]
The limits are per-user.
Each user in the same user namespace
can create namespaces up to the defined limit.
.IP \[bu]
The limits apply to all users, including UID 0.
.IP \[bu]
These limits apply in addition to any other per-namespace
limits (such as those for PID and user namespaces) that may be enforced.
.IP \[bu]
Upon encountering these limits,
.BR clone (2)
and
.BR unshare (2)
fail with the error
.BR ENOSPC .
.IP \[bu]
For the initial user namespace,
the default value in each of these files is half the limit on the number
of threads that may be created
.RI ( /proc/sys/kernel/threads\-max ).
In all descendant user namespaces, the default value in each file is
.BR MAXINT .
.IP \[bu]
When a namespace is created, the object is also accounted
against ancestor namespaces.
More precisely:
.RS
.IP \[bu] 3
Each user namespace has a creator UID.
.IP \[bu]
When a namespace is created,
it is accounted against the creator UIDs in each of the
ancestor user namespaces,
and the kernel ensures that the corresponding namespace limit
for the creator UID in the ancestor namespace is not exceeded.
.IP \[bu]
The aforementioned point ensures that creating a new user namespace
cannot be used as a means to escape the limits in force
in the current user namespace.
.RE
.\"
.SS Namespace lifetime
Absent any other factors,
a namespace is automatically torn down when the last process in
the namespace terminates or leaves the namespace.
However, there are a number of other factors that may pin
a namespace into existence even though it has no member processes.
These factors include the following:
.IP \[bu] 3
An open file descriptor or a bind mount exists for the corresponding
.IR /proc/ pid /ns/*
file.
.IP \[bu]
The namespace is hierarchical (i.e., a PID or user namespace),
and has a child namespace.
.IP \[bu]
It is a user namespace that owns one or more nonuser namespaces.
.IP \[bu]
It is a PID namespace,
and there is a process that refers to the namespace via a
.IR /proc/ pid /ns/pid_for_children
symbolic link.
.IP \[bu]
It is a time namespace,
and there is a process that refers to the namespace via a
.IR /proc/ pid /ns/time_for_children
symbolic link.
.IP \[bu]
It is an IPC namespace, and a corresponding mount of an
.I mqueue
filesystem (see
.BR mq_overview (7))
refers to this namespace.
.IP \[bu]
It is a PID namespace, and a corresponding mount of a
.BR proc (5)
filesystem refers to this namespace.
.SH EXAMPLES
See
.BR clone (2)
and
.BR user_namespaces (7).
.SH SEE ALSO
.BR nsenter (1),
.BR readlink (1),
.BR unshare (1),
.BR clone (2),
.BR ioctl_nsfs (2),
.BR setns (2),
.BR unshare (2),
.BR proc (5),
.BR capabilities (7),
.BR cgroup_namespaces (7),
.BR cgroups (7),
.BR credentials (7),
.BR ipc_namespaces (7),
.BR network_namespaces (7),
.BR pid_namespaces (7),
.BR user_namespaces (7),
.BR uts_namespaces (7),
.BR lsns (8),
.BR switch_root (8)