check: detect large omaps with `ceph health detail`

ceph / ceph-medic

find common issues in ceph clusters

MIT License

22 stars 18 forks source link

check: detect large omaps with `ceph health detail` #133

Open alfredodeza opened 5 years ago

alfredodeza commented 5 years ago

Todo: Find out what exactly is considered large so that a proper check can be created

jeanchlopez commented 5 years ago

If a bucket index shard has more than 100,000 index entries it should be considered large given the best practice recommends to use this as a limit.

alfredodeza commented 5 years ago

Is this something that is now being reported via ceph health detail ? I thought this was through the radosgw-admin tool (which unfortunately can hang...)

jeanchlopez commented 5 years ago

Not sure it is reported in 3.x versions. For sure present Mimic and onward. I'll double check when this made it into the code and confirm.

vasukulkarni commented 5 years ago

Nice to see these checks in ceph-medic!