General

Profile

Niklas Hambuechen

  • Registered on: 03/01/2018
  • Last connection: 12/13/2023

Issues

Activity

Today

03:36 AM Ceph Bug #65199: autoscaler: Scale PGs based on number of objects
Aside:
It also feels wrong that the inode information shows up as "0 B". It looks like storage is gone missing ine...
12:24 AM bluestore Bug #53899: bluefs _allocate allocation failed - BlueFS.cc: 2768: ceph_abort_msg("bluefs enospc")
Small addition: `ceph osd status` shows the following:...
12:16 AM bluestore Bug #53899: bluefs _allocate allocation failed - BlueFS.cc: 2768: ceph_abort_msg("bluefs enospc")
@Igor
It seems I hit this bug with a cluster initially that had initially deployed 16.2.7 and then upgraded to 18....

03/28/2024

12:42 PM Ceph Bug #65199 (New): autoscaler: Scale PGs based on number of objects
Ceph's autoscaler scales PGs based on Bytes stored. It seemingly ignores number of objects. This creates problems for...

03/20/2024

03:09 AM CephFS Bug #64852: MDS hangs on "joining batch getattr" when client does statx
> In upstream we have fixed it since v6.8-rc5.
All clear. I will best be able to test that in ~2-3 months, when my...

03/13/2024

04:03 AM CephFS Bug #64852: MDS hangs on "joining batch getattr" when client does statx
@Xiubo Li Could you shortly summarise them for me? For the second one, it's marked as Resolved but there seems to be ...

03/12/2024

04:58 AM CephFS Bug #61947: mds: enforce a limit on the size of a session in the sessionmap
Potentially related issue:
* https://tracker.ceph.com/issues/64852 - MDS hangs on "joining batch getattr" when cli...
04:58 AM CephFS Bug #63364: MDS_CLIENT_OLDEST_TID: 15 clients failing to advance oldest client/flush tid
Potentially related issue:
* https://tracker.ceph.com/issues/64852 - MDS hangs on "joining batch getattr" when cli...
04:57 AM CephFS Bug #64852: MDS hangs on "joining batch getattr" when client does statx
Are these possibly related?
* https://tracker.ceph.com/issues/63364 - MDS_CLIENT_OLDEST_TID: 15 clients failing to...
04:55 AM CephFS Bug #64852: MDS hangs on "joining batch getattr" when client does statx
Some logs from `/var/log/ceph/ceph-mds.mycluster-node-4.log` to show that the problematic op hung for multiple hours:...

Also available in: Atom