Hello. I have a few issues with my ceph cluster: - RGWs have disappeared from management (console does not register any RGWs) despite showing 4 services deployed and processes running; - All object buckets not accessible / manageable; - Console showing some of my pools are “updating” – its been like this for a few days; What was done: - Expanded a 30 OSD / 3 node cluster to 60 OSDs across 6 nodes; - Changed the failure domain of our crush rules from OSD to host; - Increased PG and PGP of main data pools to reflect additional cluster OSDs and maintain a ~ 100 PG / OSD target; - Attempted to redeploy / re-create service for RGWs / remove SSL config; - A single OSD with very high ECC errors was found and preemptively removed from the cluster. The cluster took a few days rebalancing itself, and impression was it would be done by now. It’s not healthy, and as per above, RGWs are no longer manageable – not sure where to start troubleshooting this as I’ve never encountered such a scenario before. Cluster specs: - 6 OSD nodes (10 OSDs each); - 5 Monitors; - 2 Managers; - 5 MDS; - 4 RGWs; - Quincy 17.2.5; - _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |