[CRIU] CRIU + Docker: can't restore into a new container
Manuel Rodríguez Pascual
manuel.rodriguez.pascual at gmail.com
Wed Aug 10 05:10:37 PDT 2016
Hi Andrei,
Thanks for your help!
Sorry, I messed it up a little bit when copy/pasting from different
tests on my previous mail, that's why both "looper2" and
"looper-force" appear . Anyway, I discovered that my problem was not
indicating the output folder when performing a checkpoint, so when I
tried to restore indicating a given folder, it was in fact empty. It
is now solved and working :)
Performing some quick tests, I have realised that if the folder
specified with "--image-dir" flag does not exist or does not contain
the necessary restore information, docker daemon will send a
"CMDNOTFOUND" error. I fully understand that this is a development
version and developers have many different things to do, but maybe
changing this error to something a bit more understandable from the
user's point of view ("FOLDERNOTFOUND" or something) would help to
understand what's going underneath.
Again, thanks for your help. Best regards,
Manuel
2016-08-10 11:46 GMT+02:00 Andrei Tuicu <andrei.tuicu at gmail.com>:
> Hi Manuel,
>
> You need to create the "looper2" container from the image container, before
> restoring the criu image on top of it.
>
> Best regards,
> Andrei
>
>
> Pe 10 aug. 2016 12:38, "Manuel Rodríguez Pascual"
> <manuel.rodriguez.pascual at gmail.com> a scris:
>>
>> Hi,
>>
>> I am a newbie with CRIU and Docker. I am exploring the possibilities
>> of docker migration with CRIU. However, I cannot restart the container
>> with a different name, as stated in the documentation. I was wondering
>> whether you could help me and point what I am doing wrong.
>>
>> Below there are some tests. I have included information on the
>> versions of CRIU and docker, a successful test on
>> checkpointing/restarting an image, and a failed test on restarting it
>> with a different name.
>>
>> Thanks for your help,
>>
>>
>> Manuel
>>
>>
>> ----
>> ----
>> system info
>>
>> $ uname -r
>> 4.7.0-1.el7.elrepo.x86_64
>>
>> $ criu --version
>> Version: 2.4
>>
>> $ docker --version
>> Docker version 1.10.0-dev, build cec56b2, experimental
>>
>>
>> ---
>> ---
>> Successful experiment
>>
>>
>> [root at tecno15 tests]# docker run -d --name looper1 --security-opt
>> seccomp:unconfined busybox /bin/sh -c 'i=0; while true; do echo $i;
>> i=$(expr $i + 1); sleep 1; done'
>> b6ade11e3202a86ac540240bca871b48c09fffa8d2432101ab3e36d75882437e
>>
>>
>> [root at tecno15 tests]# docker run -d --name looper1 --security-opt
>> seccomp:unconfined busybox /bin/sh -c 'i=0; while true; do echo $i;
>> i=$(expr $i + 1); sleep 1; done'
>> b6ade11e3202a86ac540240bca871b48c09fffa8d2432101ab3e36d75882437e
>>
>> [root at tecno15 tests]# docker ps
>> CONTAINER ID IMAGE COMMAND
>> CREATED STATUS PORTS NAMES
>> b6ade11e3202 busybox "/bin/sh -c 'i=0; whi" 16
>> seconds ago Up 14 seconds looper1
>>
>> [root at tecno15 tests]# docker logs looper1
>> 0
>> 1
>> 2
>> 3
>> 4
>> 5
>>
>> [root at tecno15 tests]# docker checkpoint looper1
>> looper1
>>
>> [root at tecno15 tests]# docker ps
>> CONTAINER ID IMAGE COMMAND CREATED
>> STATUS PORTS NAMES
>>
>> [root at tecno15 tests]# docker restore looper1
>> looper1
>>
>> [root at tecno15 tests]# docker ps
>> CONTAINER ID IMAGE COMMAND
>> CREATED STATUS PORTS NAMES
>> b6ade11e3202 busybox "/bin/sh -c 'i=0; whi" About
>> a minute ago Up 1 seconds looper1
>>
>> [root at tecno15 tests]# docker logs looper1
>> 0
>> 1
>> 2
>> 3
>> 4
>> 5
>> 6
>> 7
>> 8
>> 9
>> 10
>> 36
>> 37
>> 38
>> 39
>> 40
>>
>> ---
>> ---
>> Failed experiment
>>
>>
>> [root at tecno15 tests]# docker create --name looper-force --security-opt
>> seccomp:unconfined busybox /bin/sh -c 'i=0; while true; do echo $i;
>> i=$(expr $i + 1); sleep 1; done'
>> d0ce905ea76fd63d544824e21cea7d5a7f8be953501b3551b0197f5b4e952c6b
>>
>> [root at tecno15 tests]# docker ps -a
>> CONTAINER ID IMAGE COMMAND
>> CREATED STATUS PORTS NAMES
>> d0ce905ea76f busybox "/bin/sh -c 'i=0; whi" 11
>> seconds ago Created looper-force
>> b6ade11e3202 busybox "/bin/sh -c 'i=0; whi" 5
>> minutes ago Up 3 minutes looper1
>>
>> [root at tecno15 tests]# docker checkpoint looper1
>> looper1
>>
>> [root at tecno15 tests]# docker ps -a
>> CONTAINER ID IMAGE COMMAND
>> CREATED STATUS PORTS
>> NAMES
>> d0ce905ea76f busybox "/bin/sh -c 'i=0; whi" 51
>> seconds ago Created
>> looper-force
>> b6ade11e3202 busybox "/bin/sh -c 'i=0; whi" 5
>> minutes ago Checkpointed 24 seconds ago
>> looper1
>>
>> [root at tecno15 tests]# docker ps
>> CONTAINER ID IMAGE COMMAND CREATED
>> STATUS PORTS NAMES
>>
>> [root at tecno15 tests]# docker restore --force=true
>> --image-dir=/home/supermanue/Downloads/tests/checkpoint looper2
>> Error response from daemon: No such container: looper2
>> Error: failed to restore one or more containers
>>
>>
>> ---
>> ---
>>
>> Output from docker daemon (just last command)
>>
>> DEBU[76347] Calling POST /v1.22/containers/looper-force/restore
>> DEBU[76347] POST /v1.22/containers/looper-force/restore?force=1
>> DEBU[76347] form data:
>>
>> {"ImagesDirectory":"/home/supermanue/Downloads/tests/checkpoint","LeaveRunning":false,"WorkDirectory":""}
>> DEBU[76347] devmapper:
>>
>> activateDeviceIfNeeded(4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76348] container mounted via layerStore:
>>
>> /var/lib/docker/0.0/devicemapper/mnt/4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219/rootfs
>> DEBU[76348] Assigning addresses for endpoint looper-force's interface
>> on network bridge
>> DEBU[76348] RequestAddress(LocalDefault/172.18.0.0/16, <nil>, map[])
>> DEBU[76348] Assigning addresses for endpoint looper-force's interface
>> on network bridge
>> DEBU[76348] Releasing addresses for endpoint looper-force's interface
>> on network bridge
>> DEBU[76348] ReleaseAddress(LocalDefault/172.18.0.0/16, 172.18.0.4)
>> DEBU[76348] devmapper:
>>
>> UnmountDevice(hash=4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76348] devmapper:
>>
>> Unmount(/var/lib/docker/0.0/devicemapper/mnt/4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76348] devmapper: Unmount done
>> DEBU[76348] devmapper:
>>
>> deactivateDevice(4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76348] devmapper: removeDevice
>>
>> START(docker-253:1-538129286-4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76368] devmapper: removeDevice
>>
>> END(docker-253:1-538129286-4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76368] devmapper: deactivateDevice
>> END(4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> DEBU[76368] devmapper:
>>
>> UnmountDevice(hash=4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219)
>> END
>> ERRO[76368] devmapper: Error unmounting device
>> 4be7e83c48a3a448e1cc1a106b65172c9ac65c5e8385bcbcc58e69d65c1b9219:
>> Device is Busy
>> ERRO[76368] Error unmounting container
>> d0ce905ea76fd63d544824e21cea7d5a7f8be953501b3551b0197f5b4e952c6b:
>> Device is Busy
>> WARN[76368] failed to cleanup ipc mounts:
>> failed to umount
>>
>> /var/lib/docker/0.0/containers/d0ce905ea76fd63d544824e21cea7d5a7f8be953501b3551b0197f5b4e952c6b/shm:
>> invalid argument
>> failed to umount
>>
>> /var/lib/docker/0.0/containers/d0ce905ea76fd63d544824e21cea7d5a7f8be953501b3551b0197f5b4e952c6b/mqueue:
>> invalid argument
>> ERRO[76368] Error unmounting container
>> d0ce905ea76fd63d544824e21cea7d5a7f8be953501b3551b0197f5b4e952c6b: not
>> mounted
>> ERRO[76368] Handler for POST /v1.22/containers/looper-force/restore
>> returned error: Cannot restore container looper-force: CMDNOTFOUND
>> _______________________________________________
>> CRIU mailing list
>> CRIU at openvz.org
>> https://lists.openvz.org/mailman/listinfo/criu
More information about the CRIU
mailing list