W/o this patch, haproxy daemon under load never can
be stopped gracefully and it also may exceed given stop/monitor
timeouts ending up being restarted by pacemaker.
The solution is:
* increase haproxy resource monitor/stop timeout and interval
* fix stop action for OCF to not rely on kill command exit codes
* wrap kill in retries and issue a SIGKILL only if there are no more
retries left for gracefull termination
* add debug parameter for haproxy RA OCF script and provide more
logging as well
* set debug for haproxy RA, if global debug is enabled
Reviewed: https:/ /review. openstack. org/159188 /git.openstack. org/cgit/ stackforge/ fuel-library/ commit/ ?id=7db84c49018 1b48a37f1bb5377 3facb77458daeb
Committed: https:/
Submitter: Jenkins
Branch: master
commit 7db84c490181b48 a37f1bb53773fac b77458daeb
Author: Bogdan Dobrelya <email address hidden>
Date: Wed Feb 25 17:53:22 2015 +0100
Adjust stop/monitor and fix kill for ns_haproxy
W/o this patch, haproxy daemon under load never can
be stopped gracefully and it also may exceed given stop/monitor
timeouts ending up being restarted by pacemaker.
The solution is:
* increase haproxy resource monitor/stop timeout and interval
* fix stop action for OCF to not rely on kill command exit codes
* wrap kill in retries and issue a SIGKILL only if there are no more
retries left for gracefull termination
* add debug parameter for haproxy RA OCF script and provide more
logging as well
* set debug for haproxy RA, if global debug is enabled
Closes-bug: #1424959
Related-bug: #1425579
Change-Id: I7ab6dc2341821c 3b82ef3d3ac63b6 4a5a9958fa9
Signed-off-by: Bogdan Dobrelya <email address hidden>