> Most are from not scrubbed since end of August … That is lucky! On an inherited Ceph instance I had got most of them unscrubbed for 1-2 years. :-) The usual reason for delays in scrubbing are insufficient IOPS (both) and even insufficient bandwidth (deep scrubbing). Scrubbing like balancing and replicating etc. are very heavy background "admin" operations that consume a lot of IOPS (and bandwidth) and many Ceph instances that I have seen can barely cope with the user workload so the background "admin" workload gets very much delayed as a result. One useful check is to increase the priority of background "admin" operations so they get done without delay and if users start complaining that is indeed the situation :-). _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |