Article ID: 123021, created on Sep 28, 2014, last review on Jun 17, 2016

  • Applies to:
  • Virtuozzo 6.0

Symptoms

Migration of a running container in VZFS4 layout without conversion does not work in Parallels Cloud Server 6.0 update 7 if the system runs with an old kernel (e.g. 2.6.32-042stab076.5). The operation fails with the similar output:

Can't move/copy CT#1234 -> CT#1234, [], [] : /usr/sbin/vzctl exited with code 50

The verbose log looks like this:

~# vzmigrate --online -vvv -r no --keep-dst TAR.GET.SER.VER 1234
...
1411398717: begin stage : startTrack
1411398717: Tracker started
1411398717: tracker pid 7442, file /vz/tmp//track_file.tTeHk0
1411398717: end stage : startTrack
1411398717: Copy private area '/vz/private/1234'
1411398717: 1411398717: Command : first
...
1411398727: begin stage : stopTrack
1411398727: waitpid() : No child processes
1411398727: Can't do tracking, do slow migrate : waitpid() : No child processes
1411398727: Syncing tracked files from '/vz/private/1234'
1411398727: Send command: second_checksum
1411398727: 1411398727: Command : second_checksum
1411398727: 1411398727: /usr/local/share/vzlinmigrate/rsync -E -q -a -H -H --sparse --numeric-ids --timeout 3600 --delete --server -c . /vz/private/1234/fs 
1411398727: /usr/local/share/vzlinmigrate/rsync -E -q -a -H -H --sparse --numeric-ids --timeout 3600 --fdin 8 --fdout 7 --write-timeout 3600 --delete -c /vz/private/1234/ 0.0.0.0:/dummy 
1411398730: done
1411398730: begin stage : preFinalStage
...
1411398730: 1411398730: Undumping CT#1234 ...
1411398730: 1411398730: /usr/sbin/vzctl --skiplock --skipowner --ignore-ha-cluster restore 1234 --undump --skip_arpdetect --dumpfile /vz/dump/dumpfile.0e1Hdq --context bde32 
1411398731: 1411398731: vzctl : Cannot mount: /vz/template:/vz/private/1234/fs /vz/root/1234: No such file or directory
1411398761: 1411398761: /usr/sbin/vzctl exited with code 50
1411398764: 1411398764: /usr/sbin/vzctl exited with code 50
1411398764: 1411398764: error [-52] : /usr/sbin/vzctl exited with code 50
1411398764: /usr/sbin/vzctl exited with code 50
1411398764: can not undump CT#1234 : /usr/sbin/vzctl exited with code 50
This CT can't be migrated online at the moment. Try offline migration.
1411398764: Dumpfile /vz/dump/dumpfile.aFGYrt.saved saved
1411398764: Can't move/copy CT#1234 -> CT#1234, [], [] : /usr/sbin/vzctl exited with code 50
1411398764: cleaning : resume CT#1234
...

Cause

Due to unexpected termination of the tracker process, the migration copies the data from /vz/private/CTID to /vz/private/CTID/fs and the resulting container's layout is broken and the private area cannot be mounted.

This issue has been reported as the internal request PSBM-29087.

Resolution

The fix is included in PCS 6 update 8.

Additional

Search Words

Can't move/copy

vzctl exited with code 50

/usr/sbin/vzctl exited with code 50

c62e8726973f80975db0531f1ed5c6a2 2897d76d56d2010f4e3a28f864d69223 0dd5b9380c7d4884d77587f3eb0fa8ef

Email subscription for changes to this article
Save as PDF