Bug #5192

oned lock file not cleaned on a daemon failure

Added by Vlastimil Holer about 4 years ago. Updated about 4 years ago.

Status:ClosedStart date:06/13/2017
Priority:NormalDue date:
Assignee:Vlastimil Holer% Done:

0%

Category:Packaging
Target version:Release 5.4
Resolution: Pull request:
Affected Versions:OpenNebula 5.2

Description

There are some cases when oned fails to stop correctly and the lock file /var/lock/one/one is left on the disk. Same situation is if the oned is killed by force (e.g. by oomkiller). Left lock file prevents future starts and lock must be cleaned manually.

# pkill -9 oned
# ls /var/lock/one/one 
/var/lock/one/one
# systemctl start opennebula
# pgrep oned

With log content:

systemd[1]: Starting OpenNebula Cloud Controller Daemon...
systemd[1]: Started OpenNebula Cloud Controller Daemon.
oned[27271]: Error: Cannot start oned, opening lock file /var/lock/one/one
systemd[1]: opennebula.service: main process exited, code=exited, status=255/n/a
systemd[1]: opennebula.service: control process exited, code=exited status=1
systemd[1]: Stopped OpenNebula Cloud Controller Daemon.
systemd[1]: Unit opennebula.service entered failed state.
thunder systemd[1]: opennebula.service failed.

Associated revisions

Revision 1f00586c
Added by Vlastimil Holer about 4 years ago

B #5192: Cleanup oned lockfile on systemd service stop/fail (#340)

History

#1 Updated by Vlastimil Holer about 4 years ago

  • Status changed from Pending to Closed
  • Assignee set to Vlastimil Holer

Also available in: Atom PDF