2024-09-23T02:52:32.943Z

<cz tan> thanks for your replay, i will try again☺️

2024-09-23T09:56:10.875Z

<cz tan> hi, The logs of the Ceph Mon\MGR are all written to /var/log/messages, and they are not generated separately. Is there a parameter that needs to be set to resolve this issue?

2024-09-23T10:01:12.520Z

<Michal Nasiadka> Hello there - I'm quite unaware how the backports in Ceph land work - should I do the backport, or is that something that will be done automatically by some bot? (re: <https://tracker.ceph.com/issues/68028#change-276875>)

2024-09-23T11:45:21.913Z

<Eugen Block> Check out the docs. You can disable log-to-journal and enable log-to-file instead. I always do that as well, I like the old, pre-cephadm behaviour in that regard.

2024-09-23T12:38:42.325Z

<Ken Carlile> so when using
```radosgw-admin bucket rm --purge-objects --bypass-gc```
on a pool that has versioning enabled, I am becoming suspicious that it's just creating more and more versions rather than torching things. they may not be _valid_ versions, but my rados object count is still extremely high while the rgw bucket object count is quite low. The rados object count (from `ceph df` is somewhere around 44M, while the bucket object count from `radosgw-admin bucket stats --bucket <bucketname>` is 20. Not 20M. 20.

2024-09-23T12:39:43.005Z

<Ken Carlile> used space on the .data pool is around 12 GB, while the .index pool is 145GB. There are no other rgw buckets in existence; nothing else in these two pools.

2024-09-23T12:41:15.072Z

<Ken Carlile> the bucket rm command is spewing errors:
```2024-09-23T07:58:03.336-0400 7f7aaa87b880  0 ERROR: bucket shard callback failed. obj=<path/to/rgwobject>. ret=(2) No such file or directory```