CephXattrSetLatency

Ceph Metadata Server (MDS) is experiencing high latency when setting extended attributes (xattr) on files. The alert fires when the average setxattr operation latency exceeds 30 milliseconds over a 5-minute period.

Impact: Medium

Diagnosis

  1. Check MDS status and health.
  2. Navigate to the OpenShift console.
  3. Access the Workloads section and select the MDS pod.
  4. Click the Metrics tab to review CPU and memory usage.
  5. Check the network issues as network latency between MDS and OSDs can cause slow metadata operations.
  6. Check OSD performance, as slow OSD operations can cascade into MDS latency issues.

Mitigation

You need to increase MDS CPU resources, MDS cache memory, scale out with multiple active MDS, use faster metadata storage, address network issues, address underlying OSD issue, or restart MDS. For more information, see the Description and Runbook section of the alert.