[CRIU] Checkpoint/Restore LXC container with CRIU on a raspberry pi 2

alex vk avankemp at gmail.com
Thu Jun 9 08:53:26 PDT 2016


Hi everyone,

I'm now trying to checkpoint and restore an LXC container on a raspberry PI
2. (Criu 2.0 is working fine with the kernel 3.18, and LXC 2.0 works as
well on the RPI2)
Checkpointing a container apparently succeeds after disabling seccomp and
adding these lines to the container config (according to the wiki):

lxc.seccomp =
# hax for criu
lxc.console = none
lxc.tty = 0
lxc.cgroup.devices.deny = c 5:1 rwm

However restoring fails with this message : "lxc-checkpoint: criu.c:
exec_criu: 305 lxc.console configured on source host but not target"

I'read on this post (
https://lists.linuxcontainers.org/pipermail/lxc-devel/2016-March/013856.html)
not to pass the --ext-mount-map option when console=none.
In this case, I have another error : "1: Error (mount.c:3381): mnt: New
root and old root are the same"

Interestingly, I have another error when trying to restore, if i run all
the container creation, start, dump and restore with sudo (not directly in
root):

pie: 1: Error (pie/restorer.c:888): Can't restore ticks/time for timer>
pie: 1: fd - 0
pie: 1: Error (pie/restorer.c:1408): Can't restore timerfd -22

I've put the full log of dumping and restoring, as well as the container
config attached below, if anyone has any ideas ?

Thanks for your help,
Alex


----------------------------------------------------------------------------------------------------
Here is what i get from the restore :

root at pi-desktop:/home/pi/checkpoint# lxc-checkpoint -r -D
/home/pi/checkpoint/ -n test-container --logfile=debug_restore
--logpriority=DEBUG
lxc-checkpoint: criu.c: do_restore: 705 criu process exited 1, output:
lxc-checkpoint: criu.c: exec_criu: 305 lxc.console configured on source
host but not target


lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/systemd//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/blkio//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/freezer//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/debug//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/cpu//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/devices//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/perf_event//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/cpuset//lxc/test-container
lxc-checkpoint: cgfsng.c: recursive_destroy: 983 Error destroying
/sys/fs/cgroup/net_cls//lxc/test-container
lxc-checkpoint: criu.c: __criu_restore: 953 restore process died
Restoring test-container failed.

----------------------------------------------------------------------------------------------------
Here is the log of the dumping created with the lxc-checkpoint -s
--logfile=debug_dump :
root at pi-desktop:/home/pi/checkpoint# lxc-checkpoint -s -D
/home/pi/checkpoint/ -n test-container --logfile=debug_dump
--logpriority=DEBUG
root at pi-desktop:/home/pi/checkpoint# cat debug_dump
 lxc-checkpoint 20160609165234.776 WARN     lxc_confile -
confile.c:config_pivotdir:1879 - lxc.pivotdir is ignored.  It will soon
become an error.
 lxc-checkpoint 20160609165234.777 WARN     lxc_confile -
confile.c:config_personality:1117 - unsupported personality 'armhf'
 lxc-checkpoint 20160609165234.784 DEBUG    lxc_commands -
commands.c:lxc_cmd_get_state:579 - 'test-container' is in 'RUNNING' state
 lxc-checkpoint 20160609165234.799 INFO     lxc_criu - criu.c:exec_criu:387
- execing: /usr/local/sbin/criu dump --tcp-established --file-locks
--link-remap --manage-cgroups --ext-mount-map auto
--enable-external-sharing --enable-external-masters --enable-fs hugetlbfs
--enable-fs tracefs -D /home/pi/checkpoint/ -o
/home/pi/checkpoint//dump.log -t 4617 --freeze-cgroup
/sys/fs/cgroup/freezer///lxc/test-container --force-irmap
-------------------------------------------------------------------------------------------

And here is the dump.log :
root at pi-desktop:/home/pi/checkpoint# cat dump.log
Warn  (autofs.c:77): Failed to find pipe_ino option (old kernel?)
Detected FS_EVENT_ON_CHILD bit in mask (will be ignored on restore)
tar: ./systemd/journal/socket: socket ignored
tar: ./systemd/journal/stdout: socket ignored
tar: ./systemd/journal/syslog: socket ignored
tar: ./systemd/journal/dev-log: socket ignored
tar: ./systemd/private: socket ignored
tar: ./systemd/notify: socket ignored

----------------------------------------------------------------------------------------------------

Here is the log created with lxc-checkpoint -r --logfile :

root at pi-desktop:/home/pi/checkpoint# cat debug_restore
 lxc-checkpoint 20160609165308.055 WARN     lxc_confile -
confile.c:config_pivotdir:1879 - lxc.pivotdir is ignored.  It will soon
become an error.
 lxc-checkpoint 20160609165308.056 WARN     lxc_confile -
confile.c:config_personality:1117 - unsupported personality 'armhf'
 lxc-checkpoint 20160609165308.075 INFO     lxc_lsm - lsm/lsm.c:lsm_init:48
- LSM security driver nop
 lxc-checkpoint 20160609165308.076 DEBUG    lxc_start -
start.c:setup_signal_fd:289 - sigchild handler set
 lxc-checkpoint 20160609165308.076 INFO     lxc_start -
start.c:lxc_init:488 - 'test-container' is initialized
 lxc-checkpoint 20160609165308.076 INFO     lxc_cgroup -
cgroup.c:cgroup_init:68 - cgroup driver cgroupfs-ng initing for
test-container
 lxc-checkpoint 20160609165308.087 ERROR    lxc_criu - criu.c:exec_criu:305
- lxc.console configured on source host but not target
 lxc-checkpoint 20160609165308.115 INFO     lxc_conf -
conf.c:run_script_argv:367 - Executing script
'/usr/share/lxcfs/lxc.reboot.hook' for container 'test-container', config
section 'lxc'
 lxc-checkpoint 20160609165308.686 ERROR    lxc_criu -
criu.c:do_restore:705 - criu process exited 1, output:
lxc-checkpoint: criu.c: exec_criu: 305 lxc.console configured on source
host but not target


 lxc-checkpoint 20160609165308.689 INFO     lxc_conf -
conf.c:run_script_argv:367 - Executing script
'/usr/share/lxcfs/lxc.reboot.hook' for container 'test-container', config
section 'lxc'
 lxc-checkpoint 20160609165309.206 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/systemd//lxc/test-container
 lxc-checkpoint 20160609165309.208 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/blkio//lxc/test-container
 lxc-checkpoint 20160609165309.209 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/freezer//lxc/test-container
 lxc-checkpoint 20160609165309.211 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/debug//lxc/test-container
 lxc-checkpoint 20160609165309.212 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/cpu//lxc/test-container
 lxc-checkpoint 20160609165309.213 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/devices//lxc/test-container
 lxc-checkpoint 20160609165309.214 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/perf_event//lxc/test-container
 lxc-checkpoint 20160609165309.215 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/cpuset//lxc/test-container
 lxc-checkpoint 20160609165309.217 ERROR    lxc_cgfsng -
cgfsng.c:recursive_destroy:983 - Error destroying
/sys/fs/cgroup/net_cls//lxc/test-container
 lxc-checkpoint 20160609165309.220 ERROR    lxc_criu -
criu.c:__criu_restore:953 - restore process died


----------------------------------------------------------------------------------------------------

The log without --ext-mount-map

/usr/local/sbin/criu restore --tcp-established --file-locks --link-remap
--manage-cgroups --ext-mount-map auto --enable-external-sharing
--enable-external-masters --enable-fs hugetlbfs --enable-fs tracefs -D
/home/pi/checkpoint -o /home/pi/checkpoint/restore.log --root
/usr/lib/arm-linux-gnueabihf/lxc --restore-detached --restore-sibling
--pidfile /tmp/filet0XGku --cgroup-root /lxc/test-container console:
--veth-pair eth0=vethINVTTV at lxcbr0

root at pi-desktop:/home/pi/checkpoint# cat restore.log
Warn  (cr-restore.c:812): Set CLONE_PARENT | CLONE_NEWPID but it might
cause restore problem,because not all kernels support such clone flags
combinations!
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
     1: Warn  (autofs.c:77): Failed to find pipe_ino option (old kernel?)
     1: Error (mount.c:3381): mnt: New root and old root are the same
Error (cr-restore.c:988): 5521 killed by signal 9: Killed
Error (cr-restore.c:1867): Restoring FAILED.

----------------------------------------------------------------------------------------------------

As a normal user, with sudo :

pi at pi-desktop:~/checkpoint$ cat debug_restore
 lxc-checkpoint 20160609171247.434 WARN     lxc_confile -
confile.c:config_pivotdir:1879 - lxc.pivotdir is ignored.  It will soon
become an error.
 lxc-checkpoint 20160609171247.435 WARN     lxc_confile -
confile.c:config_personality:1117 - unsupported personality 'armhf'
 lxc-checkpoint 20160609171247.454 INFO     lxc_lsm - lsm/lsm.c:lsm_init:48
- LSM security driver nop
 lxc-checkpoint 20160609171247.455 DEBUG    lxc_start -
start.c:setup_signal_fd:289 - sigchild handler set
 lxc-checkpoint 20160609171247.455 INFO     lxc_start -
start.c:lxc_init:488 - 'test-container' is initialized
 lxc-checkpoint 20160609171247.455 INFO     lxc_cgroup -
cgroup.c:cgroup_init:68 - cgroup driver cgroupfs-ng initing for
test-container
 lxc-checkpoint 20160609171247.456 ERROR    lxc_cgfsng -
cgfsng.c:cgfsng_create:1072 - No such file or directory - Failed to create
/sys/fs/cgroup/systemd//lxc/test-container: No such file or directory
 lxc-checkpoint 20160609171247.466 INFO     lxc_criu - criu.c:exec_criu:387
- execing: /usr/local/sbin/criu restore --tcp-established --file-locks
--link-remap --manage-cgroups --ext-mount-map auto
--enable-external-sharing --enable-external-masters --enable-fs hugetlbfs
--enable-fs tracefs -D /home/pi/checkpoint -o
/home/pi/checkpoint/restore.log --root /usr/lib/arm-linux-gnueabihf/lxc
--restore-detached --restore-sibling --pidfile /tmp/file8x1eFF
--cgroup-root /lxc/test-container-1 --ext-mount-map console: --veth-pair
eth0=veth6AHNY2 at lxcbr0
 lxc-checkpoint 20160609171248.456 ERROR    lxc_criu -
criu.c:do_restore:705 - criu process exited 1, output:


 lxc-checkpoint 20160609171248.458 INFO     lxc_conf -
conf.c:run_script_argv:367 - Executing script
'/usr/share/lxcfs/lxc.reboot.hook' for container 'test-container', config
section 'lxc'
 lxc-checkpoint 20160609171248.984 ERROR    lxc_criu -
criu.c:__criu_restore:953 - restore process died

----------------------------------

The restore.log as a normal user, with sudo :

pi at pi-desktop:~/checkpoint$ sudo cat restore.log
Warn  (cr-restore.c:812): Set CLONE_PARENT | CLONE_NEWPID but it might
cause restore problem,because not all kernels support such clone flags
combinations!
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
RTNETLINK answers: File exists
     1: Warn  (autofs.c:77): Failed to find pipe_ino option (old kernel?)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x88b4
peer 0 (name /run/systemd/notify dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x88b6
peer 0 (name /run/systemd/private dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x88ba
peer 0 (name /run/systemd/journal/dev-log dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x8285
peer 0 (name /run/systemd/journal/stdout dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x8287
peer 0 (name /run/systemd/journal/socket dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x88d0
peer 0 (name /run/systemd/journal/syslog dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x83cc
peer 0x83cb (name /run/systemd/journal/stdout dir -)
     1: Warn  (sk-unix.c:1292): sk unix: Can't unlink stale socket 0x83fc
peer 0x8dfa (name /run/systemd/journal/stdout dir -)
pie: 1: Error (pie/restorer.c:888): Can't restore ticks/time for timer>
pie: 1: fd - 0
pie: 1: Error (pie/restorer.c:1408): Can't restore timerfd -22
pie: 1: Error (pie/restorer.c:1506): Restorer fail 1
Error (cr-restore.c:985): 6401 exited, status=1
Error (cr-restore.c:1867): Restoring FAILED.


-------------------------------

Here is the config of the container : /var/lib/lxc/test-container/config

# Template used to create this container:
/usr/share/lxc/templates/lxc-ubuntu
# Parameters passed to the template:
# For additional config options, please look at lxc.container.conf(5)

# Uncomment the following line to support nesting containers:
#lxc.include = /usr/share/lxc/config/nesting.conf
# (Be aware this has security implications)


# Common configuration
lxc.include = /usr/share/lxc/config/ubuntu.common.conf

# Container specific configuration
lxc.rootfs = /var/lib/lxc/test-container/rootfs
lxc.rootfs.backend = dir
lxc.utsname = test-container
lxc.arch = armhf

# Network configuration
lxc.network.type = veth
lxc.network.link = lxcbr0
lxc.network.flags = up
lxc.network.hwaddr = 00:16:3e:d5:62:a5
lxc.seccomp =
lxc.console = none
lxc.tty = 0
lxc.cgroup.devices.deny = c 5:1 rwm
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openvz.org/pipermail/criu/attachments/20160609/7b7f6e10/attachment-0001.html>


More information about the CRIU mailing list