1crun(1) General Commands Manual crun(1)
2
6 crun - a fast and lightweight OCI runtime
7
11 crun [global options] command [command options] [arguments...]
12
16 crun is a command line program for running Linux containers that follow
17 the Open Container Initiative (OCI) format.
18
22 create Create a container. The runtime detaches from the container
23 process once the container environment is created. It is necessary to
24 successively use start for starting the container.
25 delete Remove definition for a container.
28 exec Exec a command in a running container.
31 list List known containers.
34 kill Send the specified signal to the container init process. If no
37 signal is specified, SIGTERM is used.
38 ps Show the processes running in a container.
41 run Create and immediately start a container.
44 spec Generate a configuration file.
47 start Start a container that was previously created. A container can‐
50 not be started multiple times.
51 state Output the state of a container.
54 pause Pause all the processes in the container.
57 resume Resume the processes in the container.
60 update Update container resource constraints.
63 checkpoint Checkpoint a running container using CRIU
66 restore Restore a container from a checkpoint
69
73 By default, when running as root user, crun saves its state under the
74 /run/crun directory. As unprivileged user, instead the XDG_RUNTIME_DIR
75 environment variable is honored, and the directory $XDG_RUN‐
76 TIME_DIR/crun is used. The global option --root overrides this set‐
77 ting.
78
82 --debug Produce verbose output.
83 --log=LOG-DESTINATION Define the destination for the error and warning
86 messages generated by crun. If the error happens late in the container
87 init process, when crun already stopped watching it, then it will be
88 printed to the container stderr.
89 It is specified in the form BACKEND:SPECIFIER.
92 These following backends are supported:
95 • file:PATH
98 • journald:IDENTIFIER
100 • syslog:IDENTIFIER
102 If no backend is specified, then file: is used by default.
106 --log-format=FORMAT Define the format of the log messages. It can ei‐
109 ther be text, or json. The default is text.
110 --no-pivot Use chroot(2) instead of pivot_root(2) when creating the
113 container. This option is not safe, and should be avoided.
114 --root=DIR Defines where to store the state for crun containers.
117 --systemd-cgroup Use systemd for configuring cgroups. If not speci‐
120 fied, the cgroup is created directly using the cgroupfs backend.
121 --cgroup-manager=MANAGER Specify what cgroup manager must be used.
124 Permitted values are cgroupfs, systemd and disabled.
125 -?, --help Print a help list.
128 --usage Print a short usage message.
131 -V, --version Print program version
134
137 crun [global options] create [options] CONTAINER
138 --bundle=PATH Path to the OCI bundle, by default it is the current di‐
141 rectory.
142 --config=FILE Override the configuration file to use. The default
145 value is config.json.
146 --console-socket=SOCKET Path to a UNIX socket that will receive the
149 ptmx end of the tty for the container.
150 --no-new-keyring Keep the same session key
153 --preserve-fds=N Additional number of FDs to pass into the container.
156 --pid-file=PATH Path to the file that will contain the container
159 process PID.
160
163 crun [global options] run [options] CONTAINER
164 --bundle=BUNDLE Path to the OCI bundle, by default it is the current
167 directory.
168 --config=FILE Override the configuration file to use. The default
171 value is config.json.
172 --console-socket=SOCKET Path to a UNIX socket that will receive the
175 ptmx end of the tty for the container.
176 --no-new-keyring Keep the same session key.
179 --preserve-fds=N Additional number of FDs to pass into the container.
182 --pid-file=PATH Path to the file that will contain the container
185 process PID.
186 --detach Detach the container process from the current session.
189
192 crun [global options] delete [options] CONTAINER
193 --force Delete the container even if it is still running.
196 --regex=REGEX Delete all the containers that satisfy the specified
199 regex.
200
203 crun [global options] exec [options] CONTAINER CMD
204 --apparmor=PROFILE Set the apparmor profile for the process.
207 --console-socket=SOCKET Path to a UNIX socket that will receive the
210 ptmx end of the tty for the container.
211 --cwd=PATH Set the working directory for the process to PATH.
214 --cap=CAP Specify an additional capability to add to the process.
217 --detach Detach the container process from the current session.
220 --cgroup=PATH Specify a sub-cgroup path inside the container cgroup.
223 The path must already exist in the container cgroup.
224 --env=ENV Specify an environment variable.
227 --no-new-privs Set the no new privileges value for the process.
230 --preserve-fds=N Additional number of FDs to pass into the container.
233 --process=FILE Path to a file containing the process JSON configura‐
236 tion.
237 --process-label=VALUE Set the asm process label for the process com‐
240 monly used with selinux.
241 --pid-file=PATH Path to the file that will contain the new process PID.
244 -t --tty Allocate a pseudo TTY.
247 **-u USERSPEC --user=USERSPEC Specify the user in the form UID[:GID].
250
253 crun [global options] list [options]
254 -q --quiet Show only the container ID.
257
260 crun [global options] kill [options] CONTAINER SIGNAL
261 --all Kill all the processes in the container.
264 --regex=REGEX Kill all the containers that satisfy the specified regex.
267
270 crun [global options] ps [options]
271 --format=FORMAT Specify the output format. It must be either table or
274 json. By default table is used.
275
278 crun [global options] spec [options]
279 -b DIR --bundle=DIR Path to the root of the bundle dir (default ".").
282 --rootless Generate a config.json file that is usable by an unprivi‐
285 leged user.
286
289 crun [global options] update [options] CONTAINER
290 --blkio-weight=VALUE Specifies per cgroup weight.
293 --cpu-period=VALUE CPU CFS period to be used for hardcapping.
296 --cpu-quota=VALUE CPU CFS hardcap limit.
299 --cpu-rt-period=VALUE CPU realtime period to be used for hardcapping.
302 --cpu-rt-runtime=VALUE CPU realtime hardcap limit.
305 --cpu-share=VALUE CPU shares.
308 --cpuset-cpus=VALUE CPU(s) to use.
311 --cpuset-mems=VALUE Memory node(s) to use.
314 --kernel-memory=VALUE Kernel memory limit.
317 --kernel-memory-tcp=VALUE Kernel memory limit for TCP buffer.
320 --memory=VALUE Memory limit.
323 --memory-reservation=VALUE Memory reservation or soft_limit.
326 --memory-swap=VALUE Total memory usage.
329 --pids-limit=VALUE Maximum number of pids allowed in the container.
332 -r, --resources=FILE Path to the file containing the resources to up‐
335 date.
336
339 crun [global options] checkpoint [options] CONTAINER
340 --image-path=DIR Path for saving CRIU image files
343 --work-path=DIR Path for saving work files and logs
346 --leave-running Leave the process running after checkpointing
349 --tcp-established Allow open TCP connections
352 --ext-unix-sk Allow external UNIX sockets
355 --shell-job Allow shell jobs
358 --pre-dump Only checkpoint the container's memory without stopping the
361 container. It is not possible to restore a container from a pre-dump.
362 A pre-dump always needs a final checkpoint (without --pre-dump). It is
363 possible to make as many pre-dumps as necessary. For a second pre-dump
364 or for a final checkpoint it is necessary to use --parent-path to point
365 crun (and thus CRIU) to the pre-dump.
366 --parent-path=DIR Doing multiple pre-dumps or the final checkpoint af‐
369 ter one or multiple pre-dumps requires that crun (and thus CRIU) knows
370 the location of the pre-dump. It is important to use a relative path
371 from the actual checkpoint directory specified via --image-path. It
372 will fail if an absolute path is used.
373 --manage-cgroups-mode=MODE Specify which CRIU manage cgroup mode should
376 be used. Permitted values are soft, ignore, full or strict. Default is
377 soft.
378
381 crun [global options] restore [options] CONTAINER
382 -b DIR --bundle=DIR Container bundle directory (default ".")
385 --image-path=DIR Path for saving CRIU image files
388 --work-path=DIR Path for saving work files and logs
391 --tcp-established Allow open TCP connections
394 --ext-unix Allow external UNIX sockets
397 --shell-job Allow shell jobs
400 --detach Detach from the container's process
403 --pid-file=FILE Where to write the PID of the container
406 --manage-cgroups-mode=MODE Specify which CRIU manage cgroup mode should
409 be used. Permitted values are soft, ignore, full or strict. Default is
410 soft.
411
416 If the annotation run.oci.seccomp.receiver=PATH is specified, the sec‐
417 comp listener is sent to the UNIX socket listening on the specified
418 path. It can also set with the RUN_OCI_SECCOMP_RECEIVER environment
419 variable. It is an experimental feature, and the annotation will be
420 removed once it is supported in the OCI runtime specs. It must be an
421 absolute path.
422
425 If the annotation run.oci.seccomp.plugins=PLUGIN1[:PLUGIN2]... is spec‐
426 ified, the seccomp listener fd is handled through the specified plug‐
427 ins. The plugin must either be an absolute path or a file name that is
428 looked up by ldopen(3). More information on how the lookup is per‐
429 formed are available on the ld.so(8) man page.
430
433 If the annotation run.oci.seccomp_fail_unknown_syscall is present, then
434 crun will fail when an unknown syscall is encountered in the seccomp
435 configuration.
436
439 If the annotation run.oci.seccomp_bpf_data is present, then crun ig‐
440 nores the seccomp section in the OCI configuration file and use the
441 specified data as the raw data to the seccomp(SECCOMP_SET_MODE_FILTER)
442 syscall. The data must be encoded in base64.
443 It is an experimental feature, and the annotation will be removed once
446 it is supported in the OCI runtime specs.
447
450 If the annotation run.oci.keep_original_groups is present, then crun
451 will skip the setgroups syscall that is used to either set the addi‐
452 tional groups specified in the OCI configuration, or to reset the list
453 of additional groups if none is specified.
454
457 If the annotation run.oci.systemd.force_cgroup_v1=/PATH is present,
458 then crun will override the specified mount point /PATH with a cgroup
459 v1 mount made of a single hierarchy none,name=systemd. It is useful to
460 run on a cgroup v2 system containers using older versions of systemd
461 that lack support for cgroup v2.
462 Note: Your container host has to have the cgroup v1 mount already
465 present, otherwise this will not work. If you want to run the container
466 rootless, the user it runs under has to have permissions to this mount‐
467 point.
468 For example, as root:
471 mkdir /sys/fs/cgroup/systemd
474 mount cgroup -t cgroup /sys/fs/cgroup/systemd -o none,name=systemd,xattr
475 chown -R the_user.the_user /sys/fs/cgroup/systemd
476
480 Specify the offset to be written to /proc/self/timens_offsets when cre‐
481 ating a time namespace.
482
485 Override the name for the systemd sub cgroup created under the systemd
486 scope, so the final cgroup will be like:
487 /sys/fs/cgroup/$PATH/$SUBGROUP
490 When it is set to the empty string, a sub cgroup is not created.
494 If not specified, it defaults to container on cgroup v2, and to "" on
497 cgroup v1.
498 e.g.
501 /sys/fs/cgroup//system.slice/foo-352700.scope/container
504
508 If the run.oci.systemd.subgroup annotation is specified, yet another
509 sub-cgroup is created and the container process is moved here.
510 /sys/fs/cgroup/$PATH/$SUBGROUP/$DELEGATED-CGROUP
513 The runtime doesn't apply any limit to the $DELEGATED-CGROUP sub-
517 cgroup, the runtime uses only $PATH/$SUBGROUP.
518 The container payload fully manages $DELEGATE-CGROUP, the limits ap‐
521 plied to $PATH/$SUBGROUP still applies to $DELEGATE-CGROUP.
522 Since cgroup delegation is not safe on cgroup v1, this option is sup‐
525 ported only on cgroup v2.
526
529 If the annotation run.oci.hooks.stdout is present, then crun will open
530 the specified file and use it as the stdout for the hook processes.
531 The file is opened in append mode and it is created if it doesn't al‐
532 ready exist.
533
536 If the annotation run.oci.hooks.stderr is present, then crun will open
537 the specified file and use it as the stderr for the hook processes.
538 The file is opened in append mode and it is created if it doesn't al‐
539 ready exist.
540
543 It is an experimental feature.
544 If specified, run the specified handler for execing the container. The
547 only supported value is krun. When krun is specified, the libkrun.so
548 shared object is loaded and it is used to launch the container using
549 libkrun.
550
553 If specified, run the wasm handler for container. Allows running wasm
554 workload natively. Accepts a .wasm binary as input and if .wat is pro‐
555 vided it will automatically compiled into a wasm module. Stdout of
556 wasm module is relayed back via crun.
557
560 If the tmpcopyup option is specified for a tmpfs, then the path that is
561 shadowed by the tmpfs mount is recursively copied up to the tmpfs it‐
562 self.
563
566 If a r$FLAG mount option is specified then the flag $FLAG is set recur‐
567 sively for each children mount.
568 These flags are supported:
571 • "rro"
574 • "rrw"
576 • "rsuid"
578 • "rnosuid"
580 • "rdev"
582 • "rnodev"
584 • "rexec"
586 • "rnoexec"
588 • "rsync"
590 • "rasync"
592 • "rdirsync"
594 • "rmand"
596 • "rnomand"
598 • "ratime"
600 • "rnoatime"
602 • "rdiratime"
604 • "rnodiratime"
606 • "rrelatime"
608 • "rnorelatime"
610 • "rstrictatime"
612 • "rnostrictatime"
614
618 If the idmap option is specified then the mount is ID mapped using the
619 container target user namespace. This is an experimental feature and
620 can change at any time without notice.
621 The idmap option supports a custom mapping that can be different than
624 the user namespace used by the container.
625 The mapping can be specified after the idmap option like:
628 idmap=uids=0-1-10#10-11-10;gids=0-100-10.
629 For each triplet, the first value is the start of the backing file sys‐
632 tem IDs that are mapped to the second value on the host. The length of
633 this mapping is given in the third value.
634 Multiple ranges are separated with #.
637 These values are written to the /proc/$PID/uid_map and
640 /proc/$PID/gid_map files to create the user namespace for the idmapped
641 mount.
642 The only two options that are currently supported after idmap are uids
645 and gids.
646 When a custom mapping is specified, a new user namespace is created for
649 the idmapped mount.
650 If no option is specified, then the container user namespace is used.
653 If the specified mapping is prepended with a '@' then the mapping is
656 considered relative to the container user namespace. The host ID for
657 the mapping is changed to account for the relative position of the con‐
658 tainer user in the container user namespace.
659 For example, the mapping: uids=@1-3-10, given a configuration like
662 "uidMappings": [
665 {
666 "containerID": 0,
667 "hostID": 0,
668 "size": 1
669 },
670 {
671 "containerID": 1,
672 "hostID": 2,
673 "size": 1000
674 }
675 ]
676 will be converted to the absolute value uids=1-4-10, where 4 is calcu‐
680 lated by adding 3 (container ID in the uids= mapping) + 1 (hostID -
681 containerID for the user namespace mapping where containerID = 1 is
682 found).
683 The current implementation doesn't take into account multiple user
686 namespace ranges, so it is the caller's responsibility to split a map‐
687 ping if it overlaps multiple ranges in the user namespace. In such a
688 case, there won't be any error reported.
689
692 When running as user different than root, an user namespace is automat‐
693 ically created even if it is not specified in the config file. The
694 current user is mapped to the ID 0 in the container, and any additional
695 id specified in the files /etc/subuid and /etc/subgid is automatically
696 added starting with ID 1.
697
700 If the configuration specifies a new user namespace made of a single
701 mapping to the root user, but either the UID or the GID are set as non‐
702 zero then crun automatically creates another user namespace to map the
703 root user to the specified UID and GID.
704 It enables running unprivileged containers with UID and GID different
707 than zero, even when a single UID and GID are available, e.g. rootless
708 users on a system without newuidmap/newgidmap.
709
713 Note: cgroup v2 does not yet support control of realtime processes and
714 the cpu controller can only be enabled when all RT processes are in the
715 root cgroup. This will make crun fail while running alongside RT pro‐
716 cesses.
717 If the cgroup configuration found is for cgroup v1, crun attempts a
720 conversion when running on a cgroup v2 system.
721 These are the OCI resources currently supported with cgroup v2 and how
724 they are converted when needed from the cgroup v1 configuration.
725
728 ┌────────────┬────────────────────┬──────────────────────┬──────────────────┐
729 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
730 ├────────────┼────────────────────┼──────────────────────┼──────────────────┤
731 │limit │ memory.max │ y = x │ │
732 ├────────────┼────────────────────┼──────────────────────┼──────────────────┤
733 │swap │ memory.swap.max │ y = x - memory_limit │ the swap limit │
734 │ │ │ │ on cgroup v1 in‐ │
735 │ │ │ │ cludes the mem‐ │
736 │ │ │ │ ory usage too │
737 ├────────────┼────────────────────┼──────────────────────┼──────────────────┤
738 │reservation │ memory.low │ y = x │ │
739 └────────────┴────────────────────┴──────────────────────┴──────────────────┘
740
742 ┌────────┬────────────────────┬────────────┬─────────┐
743 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
744 ├────────┼────────────────────┼────────────┼─────────┤
745 │limit │ pids.max │ y = x │ │
746 └────────┴────────────────────┴────────────┴─────────┘
747
749 ┌────────┬────────────────────┬──────────────────┬──────────────────┐
750 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
751 ├────────┼────────────────────┼──────────────────┼──────────────────┤
752 │shares │ cpu.weight │ y = (1 + ((x - │ │
753 │ │ │ 2) * 9999) / │ │
754 │ │ │ 262142) │ │
755 ├────────┼────────────────────┼──────────────────┼──────────────────┤
756 │ │ convert from │ │ │
757 │ │ [2-262144] to │ │ │
758 │ │ [1-10000] │ │ │
759 ├────────┼────────────────────┼──────────────────┼──────────────────┤
760 │period │ cpu.max │ y = x │ period and quota │
761 │ │ │ │ are written to‐ │
762 │ │ │ │ gether │
763 ├────────┼────────────────────┼──────────────────┼──────────────────┤
764 │quota │ cpu.max │ y = x │ period and quota │
765 │ │ │ │ are written to‐ │
766 │ │ │ │ gether │
767 └────────┴────────────────────┴──────────────────┴──────────────────┘
768
770 ┌──────────────┬────────────────────┬────────────┬─────────┐
771 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
772 ├──────────────┼────────────────────┼────────────┼─────────┤
773 │weight │ io.bfq.weight │ y = x │ │
774 ├──────────────┼────────────────────┼────────────┼─────────┤
775 │weight_device │ io.bfq.weight │ y = x │ │
776 ├──────────────┼────────────────────┼────────────┼─────────┤
777 │rbps │ io.max │ y=x │ │
778 ├──────────────┼────────────────────┼────────────┼─────────┤
779 │wbps │ io.max │ y=x │ │
780 ├──────────────┼────────────────────┼────────────┼─────────┤
781 │riops │ io.max │ y=x │ │
782 ├──────────────┼────────────────────┼────────────┼─────────┤
783 │wiops │ io.max │ y=x │ │
784 └──────────────┴────────────────────┴────────────┴─────────┘
785
787 ┌────────┬────────────────────┬────────────┬─────────┐
788 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
789 ├────────┼────────────────────┼────────────┼─────────┤
790 │cpus │ cpuset.cpus │ y = x │ │
791 ├────────┼────────────────────┼────────────┼─────────┤
792 │mems │ cpuset.mems │ y = x │ │
793 └────────┴────────────────────┴────────────┴─────────┘
794
796 ┌────────────────┬────────────────────┬────────────┬─────────┐
797 │OCI (x) │ cgroup 2 value (y) │ conversion │ comment │
798 ├────────────────┼────────────────────┼────────────┼─────────┤
799 │.limit_in_bytes │ hugetlb..max │ y = x │ │
800 └────────────────┴────────────────────┴────────────┴─────────┘
801 User Commands crun(1)