StackLight

crm resource stop lma_collector is sometimes ineffective

Bug #1560910 reported by Patrick Petit
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
StackLight
Invalid
Critical
Unassigned
StackLight 0.8.0

Bug Description

We have witnessed in several in several deployments that when the lma_collector (hekad) process is wedged (stuck) the ocf script to stop hekad is ineffective. A manual kill -9 is required to effectively terminate the hekad process. This problem has been observed in LMA 0.8 and may also exist in LMA 0.9.

Revision history for this message
Denis Klepikov (dklepikov) wrote :

The main issue is that hekad process went into deadlock.
This bug is a duplicate of https://bugs.launchpad.net/lma-toolchain/+bug/1557388

Revision history for this message
Patrick Petit (patrick-michel-petit) wrote :

It is not a duplicate. This is a different bug than https://bugs.launchpad.net/lma-toolchain/+bug/1557388 .
One should not have to run kill -9 to kill hekad, crm resource stop lma_collector should do it.

Revision history for this message
Patrick Petit (patrick-michel-petit) wrote :

crm resource stop lma_collector first runs a kill -TERM followed after timeout of 15 seconds by a kill -9.

Changed in lma-toolchain:
status: New → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.