Ceph: collect cluster statistics
In order to be able to troubleshoot ceph related issue we need to collect continuously the cluster health statistics.
The following commands are proposed to be run every 5 minutes on a control storing the output in logs (format: date, command name, output):
ceph -f json status
ceph -f json health detail
ceph -f json osd dump
ceph -f json osd perf
ceph -f json osd pool stats
ceph -f json df
Logs are rotated and saved for at least one week.
Log format:
/var/
2015-06-09 00:00:58 [df] {"stats"
2015-06-09 00:00:59 [health] {"health"
2015-06-09 00:00:59 [osddump] {"epoch"
2015-06-09 00:00:59 [osdperf] {"osd_perf_
2015-06-09 00:00:59 [poolstats] [{"pool_
2015-06-09 00:00:59 [status] {"health"
These logs should be collected by fuel snapshot.
Blueprint information
- Status:
- Not started
- Approver:
- None
- Priority:
- Undefined
- Drafter:
- Mykola Golub
- Direction:
- Needs approval
- Assignee:
- Oleksiy Molchanov
- Definition:
- New
- Series goal:
- None
- Implementation:
- Unknown
- Milestone target:
- None
- Started by
- Completed by