Ceph

10/25/2021

02:08 PM rgw Bug #53029 (Resolved): radosgw-admin fails on "sync status" if a single RGW process is down

We're using ceph version 15.2.13 (c44bc49e7a57a87d84dfff2a077a2058aa2172e2) octopus (stable) in a containerized depl... David Piper

08/25/2021

03:02 PM CephFS Bug #51866: mds daemon damaged after outage

We've delayed MDS restarts with a script that waits for `active+clean` pgs first. Script is attached; it's borrowed ... David Piper

08/12/2021

10:49 AM CephFS Bug #51866: mds daemon damaged after outage

Thanks Dan, that's a convincing diagnosis. We've diverted our attention to other projects but I'll look into implemen... David Piper

07/27/2021

11:24 AM CephFS Bug #51866: mds daemon damaged after outage

MON / OSD logs attached from between 09:25:00 and 09:31:00
I'll try the MDS delay and get back to you. David Piper

09:49 AM CephFS Bug #51866 (Can't reproduce): mds daemon damaged after outage

Seen on a containerised test cluster with 3 x MON, 4 x OSD, 2 x MDS.
ceph version 15.2.13 (c44bc49e7a57a87d84dfff2... David Piper

07/13/2021

03:40 PM rgw Feature #51655 (New): No automatic alarm / recovery for unresponsive RGW

Running a multisite environment on ceph octopus 15.2.9 (container version: v5.0.10-stable-5.0-octopus-centos-8 ).
Si... David Piper

04/14/2021

02:24 PM rgw Bug #23264: Server side encryption support for s3 COPY operation

Is there any plan to fix this in upcoming releases? David Piper

02:15 PM CephFS Bug #50083: CephFS file access issues using kernel driver: file overwritten with null bytes

Hi sorry for the delay.
We were fairly panicked about this so have moved our kit onto a different kernel that does... David Piper

10:56 AM Ceph Bug #49370: No alarm if all standby MDSs have failed

Sorry - I typoed the name of this bug. This is about MGR alarms, not MDS.
So I don't think this is a duplicate of... David Piper

04/01/2021

08:32 AM CephFS Bug #50083: CephFS file access issues using kernel driver: file overwritten with null bytes

We do have a reliable repro currently yes. A file that we write to several times an hour shows clear corruption corre... David Piper