Skip to content

Logging

Ceph MON

Audit Log

By default, not changeable when using Rook Ceph, the MONs are printing out (quite a lot) logs due to audit logging.

These lines look like this (prefixed with audit):

audit 2023-07-06T12:11:37.524550+0000 mon.e (mon.2) 310555 : audit [DBG] from='admin socket' entity='admin socket' cmd='mon_status' args=[]: dispatch

Hint

The example line is coming from the Rook Ceph pod probes to ensure that the mon daemon is healthy.

Cluster Debug

Note

Due to a regression in Ceph Quincy the option mon_cluster_log_to_file_level that can control the log level is not checked, causing the logs to always appear. The ticket/ PR that changed the default value of these logs for improving "visibility" when a cluster is in trouble.

These lines can be important when troubleshooting Ceph issues though.

These lines look like this (prefixed with cluster):

cluster 2023-07-06T11:14:01.294211+0000 mgr.b (mgr.17383118) 66268 : cluster [DBG] pgmap v66351: 497 pgs: 497 active+clean; 761 GiB data, 1.5 TiB used, 2.2 TiB / 3.7 TiB avail; 114 KiB/s rd, 1.9 MiB/s wr, 211 op/s

These lines can help when debugging issues with a cluster during normal and recovery operations, though they might can be a lot depending on the cluster size.