Hello,
I've CEPH Cluster, and from a week I'm seeing strange slow perfomance in heartbeat, with high latency.
So I'm evaluating to change the default scrub schedule. is somet hing that you suggest?
Do you need some tip about it?
From WebUI, in Ceph -> OSD, you can enable noscrub and nodeepscrub flags so no scrub activity will be started. You can then monitor your latencies.
Every time I've seen latency spikes has been because of dying disk(s) or misbehaving OSD process that got solved restarting them one by one and waiting for status to be fully OK before restarting the next.
Read i.e. this [1]. Values can be set with ceph config set osd <parameter> <value>. If using Ceph >= Quincy, read about mclock [2] and how it deals with scrub operations priority.