percona-cluster had two units owning the cluster VIP
Bug #1742811 reported by
Drew Freiberger
This bug report is a duplicate of:
Bug #1740892: corosync upgrade on 2018-01-02 caused pacemaker to fail.
Edit
Remove
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Percona Cluster Charm |
New
|
Undecided
|
Unassigned |
Bug Description
We recently had a trusty-mitaka cloud running 17.02 charms have an incident caused by two of the mysql units owning the VIP at the same time. This was resolved by manually removing the VIP from one of the two servers.
It manifested as too many open files/too many connections, as the connections were bouncing back and forth between the two servers.
I'm still digging into the logs to find useful info. Recommendations of what to look for are highly welcome.
To post a comment you must log in.
This was actually caused by lp#1740892