Pacemaker Galera-MySQL restart race in OCF script
Bug #1281625 reported by
Vladimir Kuklin
This bug affects 2 people
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Fuel for OpenStack |
Fix Committed
|
High
|
Vladimir Kuklin | ||
Bug Description
there is race condition in our galera OCF mysql-wss script which can fail the deployment as it does not correctly check mysql process status.
Here:
mysql_status() {
if [ ! -e $OCF_RESKEY_pid ]; then
ocf_log $1 "MySQL is not running"
return $OCF_NOT_RUNNING;
fi
Sometimes mysqld_safe process does not create pid at the moment OCF script checks for it's existence. It turns pacemaker crazy as it goes into infinite mysql start/stop loop and fails the deployment.
| Changed in fuel: | |
| assignee: | Fuel Library Team (fuel-library) → Vladimir Kuklin (vkuklin) |
| status: | Triaged → In Progress |
Revision history for this message
| Andrew Woodward (xarses) wrote | #1 |
Revision history for this message
| OpenStack Infra (hudson-openstack) wrote Fix proposed to fuel-library (master) | #2 |
Revision history for this message
| OpenStack Infra (hudson-openstack) wrote Fix merged to fuel-library (master) | #3 |
| Changed in fuel: | |
| status: | In Progress → Fix Committed |
| tags: | added: ha |
| summary: |
- race condition in galera OCF script + Pacemaker Galera-MySQL restart race in OCF script |
Revision history for this message
| Chen Xin Jiang (jcxxin) wrote | #4 |
To post a comment you must log in.
I've seen this occur on a running deployment after doing some maintenance, it causes two mysqld_safe's to spawn if the first one doesn't start a pidfile (ubuntu)