Great, thank you both for the confirmation! -----Original Message----- From: Xiubo Li <xiubli@xxxxxxxxxx> Sent: Friday, October 21, 2022 8:43 AM To: Rishabh Dave <ridave@xxxxxxxxxx>; Edward R Huyer <erhvks@xxxxxxx> Cc: ceph-users@xxxxxxx Subject: Re: Re: MDS_CLIENT_LATE_RELEASE after setting up scheduled CephFS snapshots On 21/10/2022 19:39, Rishabh Dave wrote: > Hi Edward, > > On Wed, 19 Oct 2022 at 21:27, Edward R Huyer <erhvks@xxxxxxx> wrote: >> I recently set up scheduled snapshots on my CephFS filesystem, and ever since the cluster has been intermittently going into HEALTH_WARN with an MDS_CLIENT_LATE_RELEASE notification. >> >> Specifically: >> >> [WARN] MDS_CLIENT_LATE_RELEASE: 1 clients failing to respond to >> capability release >> >> mds.[redacted](mds.0): Client [redacted] failing to respond >> to capability release client_id: 806270628 >> >> >> I catch errors like this in the logs: >> client.785155718 isn't responding to mclientcaps(revoke), ino >> 0x10004c1abc6 pending pAsLsXsFsc issued pAsLsXsFsc, sent 3844.269321 >> seconds ago >> >> (Ignore the fact that the client numbers don't match in this case; >> one was captured before the client was rebooted, the other after. >> There's only one CephFS client and the numbers normally match.) >> >> If left alone, the issue eventually resolves itself, then comes back at some point in the future. >> >> It appears to be the same as this bug: >> https://tracker.ceph.com/issues/49434 >> Which leads here: https://tracker.ceph.com/issues/57244 >> And then this pull request: https://github.com/ceph/ceph/pull/47752 >> >> My reading is that this is simply a matter of the MDS not recognizing that the caps have, in fact, been released, and that I can safely ignore the warnings until the patch percolates down to a Pacific release. Is that right, or am I missing something significant? >> > The PR has been marked for being backported to Pacific, so the patch > will eventually end up in Pacific. In running tests for CephFS, I > haven't seen more complications from this bug than the issue you seem > to hit. This is what I see normally - "cluster [WRN] client.xxxxx > isn't responding to mclientcaps(revoke)". So ignoring it probably is > safe. I'll try to contact the PR author and ask for an opinion. Yeah, normally if it doesn't cause any stuck for your applications it should be okay. As I know it won't. And the ceph PR is still under reviewing and testing, once it gets merged I will backport it asap. Thanks! - Xiubo > - Regards, > Rishabh > >> ----- >> Edward Huyer >> Golisano College of Computing and Information Sciences Rochester >> Institute of Technology Golisano 70-2373 >> 152 Lomb Memorial Drive >> Rochester, NY 14623 >> 585-475-6651 >> erhvks@xxxxxxx<mailto:erhvks@xxxxxxx> >> >> Obligatory Legalese: >> The information transmitted, including attachments, is intended only for the person(s) or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and destroy any copies of this information. >> >> _______________________________________________ >> ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an >> email to ceph-users-leave@xxxxxxx >> > _______________________________________________ > ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an > email to ceph-users-leave@xxxxxxx > _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |