man7/namespaces.7

   1 .\" Copyright (c) 2013 by Michael Kerrisk <mtk.manpages@gmail.com>
   2 .\" and Copyright (c) 2012 by Eric W. Biederman <ebiederm@xmission.com>
   3 .\"
   4 .\" %%%LICENSE_START(VERBATIM)
   5 .\" Permission is granted to make and distribute verbatim copies of this
   6 .\" manual provided the copyright notice and this permission notice are
   7 .\" preserved on all copies.
   8 .\"
   9 .\" Permission is granted to copy and distribute modified versions of this
  10 .\" manual under the conditions for verbatim copying, provided that the
  11 .\" entire resulting derived work is distributed under the terms of a
  12 .\" permission notice identical to this one.
  13 .\"
  14 .\" Since the Linux kernel and libraries are constantly changing, this
  15 .\" manual page may be incorrect or out-of-date.  The author(s) assume no
  16 .\" responsibility for errors or omissions, or for damages resulting from
  17 .\" the use of the information contained herein.  The author(s) may not
  18 .\" have taken the same level of care in the production of this manual,
  19 .\" which is licensed free of charge, as they might when working
  20 .\" professionally.
  21 .\"
  22 .\" Formatted or processed versions of this manual, if unaccompanied by
  23 .\" the source, must acknowledge the copyright and authors of this work.
  24 .\" %%%LICENSE_END
  25 .\"
  26 .\"
  27 .TH NAMESPACES 7 2016-03-15 "Linux" "Linux Programmer's Manual"
  28 .SH NAME
  29 namespaces \- overview of Linux namespaces
  30 .SH DESCRIPTION
  31 A namespace wraps a global system resource in an abstraction that
  32 makes it appear to the processes within the namespace that they
  33 have their own isolated instance of the global resource.
  34 Changes to the global resource are visible to other processes
  35 that are members of the namespace, but are invisible to other processes.
  36 One use of namespaces is to implement containers.
  37
  38 Linux provides the following namespaces:
  39 .TS
  40 lB lB lB
  41 l lB l.
  42 Namespace       Constant        Isolates
  43 Cgroup  CLONE_NEWCGROUP Cgroup root directory
  44 IPC     CLONE_NEWIPC    System V IPC, POSIX message queues
  45 Network CLONE_NEWNET    Network devices, stacks, ports, etc.
  46 Mount   CLONE_NEWNS     Mount points
  47 PID     CLONE_NEWPID    Process IDs
  48 User    CLONE_NEWUSER   User and group IDs
  49 UTS     CLONE_NEWUTS    Hostname and NIS domain name
  50 .TE
  51
  52 This page describes the various namespaces and the associated
  53 .I /proc
  54 files, and summarizes the APIs for working with namespaces.
  55 .\"
  56 .\" ==================== The namespaces API ====================
  57 .\"
  58 .SS The namespaces API
  59 As well as various
  60 .I /proc
  61 files described below,
  62 the namespaces API includes the following system calls:
  63 .TP
  64 .BR clone (2)
  65 The
  66 .BR clone (2)
  67 system call creates a new process.
  68 If the
  69 .I flags
  70 argument of the call specifies one or more of the
  71 .B CLONE_NEW*
  72 flags listed below, then new namespaces are created for each flag,
  73 and the child process is made a member of those namespaces.
  74 (This system call also implements a number of features
  75 unrelated to namespaces.)
  76 .TP
  77 .BR setns (2)
  78 The
  79 .BR setns (2)
  80 system call allows the calling process to join an existing namespace.
  81 The namespace to join is specified via a file descriptor that refers to
  82 one of the
  83 .IR /proc/[pid]/ns
  84 files described below.
  85 .TP
  86 .BR unshare (2)
  87 The
  88 .BR unshare (2)
  89 system call moves the calling process to a new namespace.
  90 If the
  91 .I flags
  92 argument of the call specifies one or more of the
  93 .B CLONE_NEW*
  94 flags listed below, then new namespaces are created for each flag,
  95 and the calling process is made a member of those namespaces.
  96 (This system call also implements a number of features
  97 unrelated to namespaces.)
  98 .PP
  99 Creation of new namespaces using
 100 .BR clone (2)
 101 and
 102 .BR unshare (2)
 103 in most cases requires the
 104 .BR CAP_SYS_ADMIN
 105 capability.
 106 User namespaces are the exception: since Linux 3.8,
 107 no privilege is required to create a user namespace.
 108 .\"
 109 .\" ==================== The /proc/[pid]/ns/ directory ====================
 110 .\"
 111 .SS The /proc/[pid]/ns/ directory
 112 Each process has a
 113 .IR /proc/[pid]/ns/
 114 .\" See commit 6b4e306aa3dc94a0545eb9279475b1ab6209a31f
 115 subdirectory containing one entry for each namespace that
 116 supports being manipulated by
 117 .BR setns (2):
 118
 119 .in +4n
 120 .nf
 121 $ \fBls -l /proc/$$/ns\fP
 122 total 0
 123 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 cgroup -> cgroup:[4026531835]
 124 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 ipc -> ipc:[4026531839]
 125 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 mnt -> mnt:[4026531840]
 126 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 net -> net:[4026531969]
 127 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 pid -> pid:[4026531836]
 128 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 user -> user:[4026531837]
 129 lrwxrwxrwx. 1 mtk mtk 0 Apr 28 12:46 uts -> uts:[4026531838]
 130 .fi
 131 .in
 132
 133 Bind mounting (see
 134 .BR mount (2))
 135 one of the files in this directory
 136 to somewhere else in the filesystem keeps
 137 the corresponding namespace of the process specified by
 138 .I pid
 139 alive even if all processes currently in the namespace terminate.
 140
 141 Opening one of the files in this directory
 142 (or a file that is bind mounted to one of these files)
 143 returns a file handle for
 144 the corresponding namespace of the process specified by
 145 .IR pid .
 146 As long as this file descriptor remains open,
 147 the namespace will remain alive,
 148 even if all processes in the namespace terminate.
 149 The file descriptor can be passed to
 150 .BR setns (2).
 151
 152 In Linux 3.7 and earlier, these files were visible as hard links.
 153 Since Linux 3.8, they appear as symbolic links.
 154 If two processes are in the same namespace, then the inode numbers of their
 155 .IR /proc/[pid]/ns/xxx
 156 symbolic links will be the same; an application can check this using the
 157 .I stat.st_ino
 158 field returned by
 159 .BR stat (2).
 160 The content of this symbolic link is a string containing
 161 the namespace type and inode number as in the following example:
 162
 163 .in +4n
 164 .nf
 165 $ \fBreadlink /proc/$$/ns/uts\fP
 166 uts:[4026531838]
 167 .fi
 168 .in
 169
 170 The files in this subdirectory are as follows:
 171 .TP
 172 .IR /proc/[pid]/ns/cgroup " (since Linux 4.6)"
 173 This file is a handle for the cgroup namespace of the process.
 174 .TP
 175 .IR /proc/[pid]/ns/ipc " (since Linux 3.0)"
 176 This file is a handle for the IPC namespace of the process.
 177 .TP
 178 .IR /proc/[pid]/ns/mnt " (since Linux 3.8)"
 179 This file is a handle for the mount namespace of the process.
 180 .TP
 181 .IR /proc/[pid]/ns/net " (since Linux 3.0)"
 182 This file is a handle for the network namespace of the process.
 183 .TP
 184 .IR /proc/[pid]/ns/pid " (since Linux 3.8)"
 185 This file is a handle for the PID namespace of the process.
 186 .TP
 187 .IR /proc/[pid]/ns/user " (since Linux 3.8)"
 188 This file is a handle for the user namespace of the process.
 189 .TP
 190 .IR /proc/[pid]/ns/uts " (since Linux 3.0)"
 191 This file is a handle for the UTS namespace of the process.
 192 .\"
 193 .\" ==================== Cgroup namespaces ====================
 194 .\"
 195 .SS Cgroup namespaces (CLONE_NEWCGROUP)
 196 Cgroup namespaces virtualize the view of a process's cgroups as seen via
 197 .IR /proc/[pid]/cgroup
 198 (see
 199 .BR cgroups (7)).
 200
 201 Each cgroup namespace has its own set of cgroup root directories,
 202 which are the base points for the relative locations displayed in
 203 .IR /proc/[pid]/cgroup .
 204 When a process creates a new cgroup namespace using
 205 .BR clone (2)
 206 or
 207 .BR unshare (2)
 208 with the
 209 .BR CLONE_NEWCGROUP
 210 flag, then its current cgroups directories become its cgroup root directories.
 211 (This applies both for the cgroups version 1 hierarchies
 212 as well as the cgroups version 2 unified hierarchy.)
 213
 214 When viewing
 215 .IR /proc/[pid]/cgroup ,
 216 the pathname shown in the third field of each record will be
 217 relative to the reading process's cgroup root directory.
 218 If the cgroup directory of the target process lies outside
 219 the cgroup root directory for this namespace,
 220 then the pathname will show
 221 .I /..
 222 entries for each ancestor level in the cgroup hierarchy.
 223
 224 The following shell session demonstrates the effect of creating
 225 a new cgroup namespace.
 226 First, we create child cgroup in the
 227 .I freezer
 228 hierarchy, and put the shell into that cgroup:
 229
 230 .nf
 231 .in +4n
 232 $ \fBsudo mkdir \-p /sys/fs/cgroup/freezer/sub\fP
 233 $ \fBecho $$\fP                      # Show PID of this shell
 234 30655
 235 $ \fBsudo sh \-c 'echo 30655 > /sys/fs/cgroup/sub'\fP
 236 $ \fBcat /proc/self/cgroup | grep freezer\fP
 237 7:freezer:/sub
 238 .in
 239 .fi
 240
 241 Next, we use
 242 .BR unshare (1)
 243 to create a process running a shell in new user and cgroup namespaces:
 244
 245 .nf
 246 .in +4n
 247 $ \fBunshare -U -C bash\fP
 248 .in
 249 .fi
 250
 251 We then inspect the
 252 .IR /proc/[pid]/cgroup
 253 files of, respectively, the new shell process started by the
 254 .BR unshare (1)
 255 command, a process that is in the original cgroup namespace
 256 .RI ( init ,
 257 with PID 1), and a process in a sibling cgroup:
 258
 259 .nf
 260 .in +4n
 261 $ \fBcat /proc/self/cgroup | grep freezer\fP
 262 7:freezer:/
 263 $ \fBcat /proc/1/cgroup | grep freezer\fP
 264 7:freezer:/..
 265 $ \fBcat /proc/20124/cgroup | grep freezer\fP
 266 7:freezer:/../sub2
 267 .in
 268 .fi
 269
 270 The virtualization provided by cgroup namespaces can be used to prevent
 271 information leaks whereby cgroup directory paths outside of
 272 a container would otherwise be visible to processes in the container.
 273
 274 Use of cgroup namespaces requires a kernel that is configured with the
 275 .B CONFIG_CGROUPS
 276 option.
 277 .\"
 278 .\" ==================== IPC namespaces ====================
 279 .\"
 280 .SS IPC namespaces (CLONE_NEWIPC)
 281 IPC namespaces isolate certain IPC resources,
 282 namely, System V IPC objects (see
 283 .BR svipc (7))
 284 and (since Linux 2.6.30)
 285 .\" commit 7eafd7c74c3f2e67c27621b987b28397110d643f
 286 .\" https://lwn.net/Articles/312232/
 287 POSIX message queues (see
 288 .BR mq_overview (7)).
 289 The common characteristic of these IPC mechanisms is that IPC
 290 objects are identified by mechanisms other than filesystem
 291 pathnames.
 292
 293 Each IPC namespace has its own set of System V IPC identifiers and
 294 its own POSIX message queue filesystem.
 295 Objects created in an IPC namespace are visible to all other processes
 296 that are members of that namespace,
 297 but are not visible to processes in other IPC namespaces.
 298
 299 The following
 300 .I /proc
 301 interfaces are distinct in each IPC namespace:
 302 .IP * 3
 303 The POSIX message queue interfaces in
 304 .IR /proc/sys/fs/mqueue .
 305 .IP *
 306 The System V IPC interfaces in
 307 .IR /proc/sys/kernel ,
 308 namely:
 309 .IR msgmax ,
 310 .IR msgmnb  ,
 311 .IR msgmni ,
 312 .IR sem ,
 313 .IR shmall ,
 314 .IR shmmax ,
 315 .IR shmmni ,
 316 and
 317 .IR shm_rmid_forced .
 318 .IP *
 319 The System V IPC interfaces in
 320 .IR /proc/sysvipc .
 321 .PP
 322 When an IPC namespace is destroyed
 323 (i.e., when the last process that is a member of the namespace terminates),
 324 all IPC objects in the namespace are automatically destroyed.
 325
 326 Use of IPC namespaces requires a kernel that is configured with the
 327 .B CONFIG_IPC_NS
 328 option.
 329 .\"
 330 .\" ==================== Network namespaces ====================
 331 .\"
 332 .SS Network namespaces (CLONE_NEWNET)
 333 Network namespaces provide isolation of the system resources associated
 334 with networking: network devices, IPv4 and IPv6 protocol stacks,
 335 IP routing tables, firewalls, the
 336 .I /proc/net
 337 directory, the
 338 .I /sys/class/net
 339 directory, port numbers (sockets), and so on.
 340 A physical network device can live in exactly one
 341 network namespace.
 342 A virtual network device ("veth") pair provides a pipe-like abstraction
 343 .\" FIXME Add pointer to veth(4) page when it is eventually completed
 344 that can be used to create tunnels between network namespaces,
 345 and can be used to create a bridge to a physical network device
 346 in another namespace.
 347
 348 When a network namespace is freed
 349 (i.e., when the last process in the namespace terminates),
 350 its physical network devices are moved back to the
 351 initial network namespace (not to the parent of the process).
 352
 353 Use of network namespaces requires a kernel that is configured with the
 354 .B CONFIG_NET_NS
 355 option.
 356 .\"
 357 .\" ==================== Mount namespaces ====================
 358 .\"
 359 .SS Mount namespaces (CLONE_NEWNS)
 360 Mount namespaces isolate the set of filesystem mount points,
 361 meaning that processes in different mount namespaces can
 362 have different views of the filesystem hierarchy.
 363 The set of mounts in a mount namespace is modified using
 364 .BR mount (2)
 365 and
 366 .BR umount (2).
 367
 368 The
 369 .IR /proc/[pid]/mounts
 370 file (present since Linux 2.4.19)
 371 lists all the filesystems currently mounted in the
 372 process's mount namespace.
 373 The format of this file is documented in
 374 .BR fstab (5).
 375 Since kernel version 2.6.15, this file is pollable:
 376 after opening the file for reading, a change in this file
 377 (i.e., a filesystem mount or unmount) causes
 378 .BR select (2)
 379 to mark the file descriptor as readable, and
 380 .BR poll (2)
 381 and
 382 .BR epoll_wait (2)
 383 mark the file as having an error condition.
 384
 385 The
 386 .IR /proc/[pid]/mountstats
 387 file (present since Linux 2.6.17)
 388 exports information (statistics, configuration information)
 389 about the mount points in the process's mount namespace.
 390 This file is readable only by the owner of the process.
 391 Lines in this file have the form:
 392 .RS
 393 .in 12
 394 .nf
 395
 396 device /dev/sda7 mounted on /home with fstype ext3 [statistics]
 397 (       1      )            ( 2 )             (3 ) (4)
 398 .fi
 399 .in
 400
 401 The fields in each line are:
 402 .TP 5
 403 (1)
 404 The name of the mounted device
 405 (or "nodevice" if there is no corresponding device).
 406 .TP
 407 (2)
 408 The mount point within the filesystem tree.
 409 .TP
 410 (3)
 411 The filesystem type.
 412 .TP
 413 (4)
 414 Optional statistics and configuration information.
 415 Currently (as at Linux 2.6.26), only NFS filesystems export
 416 information via this field.
 417 .RE
 418 .\"
 419 .\" ==================== PID namespaces ====================
 420 .\"
 421 .SS PID namespaces (CLONE_NEWPID)
 422 See
 423 .BR pid_namespaces (7).
 424 .\"
 425 .\" ==================== User namespaces ====================
 426 .\"
 427 .SS User namespaces (CLONE_NEWUSER)
 428 See
 429 .BR user_namespaces (7).
 430 .\"
 431 .\" ==================== UTS namespaces ====================
 432 .\"
 433 .SS UTS namespaces (CLONE_NEWUTS)
 434 UTS namespaces provide isolation of two system identifiers:
 435 the hostname and the NIS domain name.
 436 These identifiers are set using
 437 .BR sethostname (2)
 438 and
 439 .BR setdomainname (2),
 440 and can be retrieved using
 441 .BR uname (2),
 442 .BR gethostname (2),
 443 and
 444 .BR getdomainname (2).
 445
 446 Use of UTS namespaces requires a kernel that is configured with the
 447 .B CONFIG_UTS_NS
 448 option.
 449 .SH CONFORMING TO
 450 Namespaces are a Linux-specific feature.
 451 .SH EXAMPLE
 452 See
 453 .BR user_namespaces (7).
 454 .SH SEE ALSO
 455 .BR lsns (1),
 456 .BR nsenter (1),
 457 .BR readlink (1),
 458 .BR unshare (1),
 459 .BR clone (2),
 460 .BR setns (2),
 461 .BR unshare (2),
 462 .BR proc (5),
 463 .BR capabilities (7),
 464 .BR cgroups (7),
 465 .BR credentials (7),
 466 .BR pid_namespaces (7),
 467 .BR user_namespaces (7),
 468 .BR switch_root (8)