Hi, If it's happening daily, around the same time, then it's possibly due to the mon scrub, which we recently found to be CPU intensive on some clusters and can cause mon elections. Do you have ScrubResult log messages in ceph log? You can also check previous days and see how long the mon scrub is taking to complete. (Time from first to last entry) Cheers, Dan On Sun, Nov 7, 2021, 9:50 AM mahnoosh shahidi <mahnooosh.shd@xxxxxxxxx> wrote: > Hi, > > We have a ceph cluster with 3 mon nodes in octopus 15.2.12. Recently, one > of our monitor nodes randomly gets out of quorum and rejoins again. The > rocksdb compaction queue of the monitor has 5 entries most of the time and > rocksdb submit sync latency is about 1 second. There isn't any problem with > mon disks. Restarting mon and mgr daemons does not help either and there is > not any special log in mon logs. > Anybody have any idea what the problem is? > > Regards > Mahnoosh > _______________________________________________ > ceph-users mailing list -- ceph-users@xxxxxxx > To unsubscribe send an email to ceph-users-leave@xxxxxxx > _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |