Have: Percona 5.6.20-rel68.0.el5
Upgraded to: Percona 5.6.21-rel69.0.el5
OS: RHEL 5.11
Replication: Percona slaves replicating off of a MySQL (community) 5.1.61 server.
Symptom: MySQL process constantly restarting. Not all of the slave servers restarted at the same time intervals.
Question: What is going wrong? Did Percona change anything between these versions that don’t handle using HugeTLB memory chunks set aside in the kernel for MySQL?
Attached is a slightly sanitized version of the *.err log
percona-error-log.txt (15 KB)
Hi,
I’m not aware of any changes regarding HugeTLB memory in 5.6.21-rel.69.0 version but still I would suggest you to upgrade from Percona 5.6.21-rel.69.0.el5 to Percona 5.6.21-rel.71.0.el5. Because, it seems there were many bugs in rel.69, which was resolved in 71.0
[url]http://www.percona.com/blog/2014/11/24/percona-server-5-6-21-70-1-now-available/[/url]
You might be affected from any of them. So please try to upgrade newer version and let us know if you still facing the issue.
Thanks for the info! I’ve upgraded Percona to the latest 5.6 version available on one of the servers and so far it seems to be doing fine. I’ll leave it running for a good long while before applying the upgrade to the rest of them, though, since the time intervals between restarts was not consistent across the servers… just in case.