On Sat, Aug 20, 2016 at 9:45 AM, Lindsay Mathieson <lindsay.mathieson@xxxxxxxxx> wrote:
On 20/08/2016 1:21 AM, David Gossage wrote:
Any issues since then? Was contemplating updating from 3.7.14 -> 3.8.2 this weekend prior to doing some work changing up the underlying brick raid levels and needing to do full heals one by one. So far has been fine on my test bed with what limited use I can put on it.
No problems at all, VM's operating normally, performance quite good.
I was wondering if my original problem was being over hasty with a "heal full" command - maybe if I had waited a few minutes it would started healing normally. Its my understanding that would do a hash comparison of all shards, which would take a long time and really thrash the CPU.
I have backups running at the moment, once they are finished I'll repeat the test and see how it does when left to its own devices.
Lindsay,
Please do "gluster volume set <volname> data-self-heal-algorithm full" to prevent diff self-heals(checksum computations on the files) which use a lot of CPU if not already. One more thing that could have lead to lot of CPU is full directory heals on .shard. Krutika recently implemented a feature called granular entry self-heal which should address this issue in future. We have throttling feature coming along in future as well to play nice with rest of the system.
--
Lindsay Mathieson
_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users
--
Pranith
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-users
