Hello, after update to Ceph Pacific 16.2.7 we’re facing issues with a lot of alerts from Alertmanager with message: OSD osd.6 on ceph-01 deviates by more than 30% from average PG count. Dashboard says 13.9 PGs per OSD, but this particular osd has only 9 PGs. For example another alert it’s saying OSD osd.19 on ceph-02.deviates by more than 30% from average PG count. osd.19 has for example 19 PGs Is this behaviour OK ? Do I need to clear CephPGImbalance alert from Alertmanager ? Or is something wrong ? We have 3 servers with 6 SSD in each and we tested it for example on version 16.2.6 and we never had this type of alerts in past. Our pools are nfs with 32 active+clean PG cephfs.CephFS.data with 32 active+clean PG cephfs.CephFS.meta with 32 active+clean PG device_health_metrics with 1 active+clean PG I all pools we are using just default PG Autoscale and replica 3. Thank you. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |