Project

General

Profile

Activity

From 12/15/2020 to 01/13/2021

01/13/2021

06:45 PM Bug #48863: cephfs-shell should allow changing all mode bits
The issue is that you can't currently set S_ISUID, S_ISGID or S_ISVTX. We should allow that within cephfs-shell.
I...
Jeff Layton
06:44 PM Bug #48863 (Resolved): cephfs-shell should allow changing all mode bits
Currently, cephfs-shell says:... Jeff Layton
04:24 PM Bug #48808 (Resolved): mon/MDSMonitor: `fs rm` is not idempotent
Patrick Donnelly
04:23 PM Bug #48834 (Resolved): qa: MDS_SLOW_METADATA_IO with osd thrasher
Patrick Donnelly
03:45 PM Backport #48859 (Resolved): nautilus: pybind/mgr/volumes: inherited snapshots should be filtered ...
https://github.com/ceph/ceph/pull/39292 Backport Bot
03:45 PM Backport #48858 (Resolved): octopus: pybind/mgr/volumes: inherited snapshots should be filtered o...
https://github.com/ceph/ceph/pull/39390 Backport Bot
03:44 PM Bug #48501 (Pending Backport): pybind/mgr/volumes: inherited snapshots should be filtered out of ...
Patrick Donnelly
01:02 PM Backport #47095: octopus: mds: provide altrenatives to increase the total cephfs subvolume snapsh...
Andras Sali wrote:
> Really looking forward to getting this fix with 15.2.9 - is there a planned release date?
@A...
Nathan Cutler
03:42 AM Bug #48778: Setting quota triggered ops storms on meta pool
Today I was able to simultaneously shut down all clients, which I did. The write operations were still happening, eve... Gerry D

01/12/2021

04:59 PM Documentation #48838 (Resolved): document ms_mode options in mount.ceph manpage
Patrick Donnelly
04:19 PM Backport #47095: octopus: mds: provide altrenatives to increase the total cephfs subvolume snapsh...
Really looking forward to getting this fix with 15.2.9 - is there a planned release date? Andras Sali
03:29 PM Bug #48770 (Resolved): qa: "Test failure: test_hole (tasks.cephfs.test_failover.TestClusterResize)"
Patrick Donnelly
01:46 PM Bug #47294 (Fix Under Review): client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Fix at: https://github.com/ceph/ceph/pull/38858
Thanks for spotting that.
Adam Emerson
12:05 AM Bug #47294: client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
That is probably the underlying cause. I'll push a patch fixing that tomorrow and then a bunch of tests and see if th... Adam Emerson
08:40 AM Bug #48778: Setting quota triggered ops storms on meta pool
Small question, why do we see client_request(mds.0:61937457 setxattr #0x10006a64656 ceph.quota caller_uid=0, caller_g... Tom Myny
05:54 AM Bug #48778: Setting quota triggered ops storms on meta pool
Thank you for your reply, Patrick.
I already suspected the requests might come from the clients, but my initial de...
Gerry D

01/11/2021

11:48 PM Bug #47294: client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
By the way, something odd I noticed today was:
https://github.com/ceph/ceph/blob/d20916964984242e513a645bd275fad89...
Patrick Donnelly
11:42 PM Bug #47294: client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Patrick Donnelly wrote:
> Adam Emerson wrote:
> > This doesn't look quite the same? Before we were having the opera...
Patrick Donnelly
07:28 PM Bug #48839 (Resolved): qa: Error: Unable to find a match: cephfs-top
My fault with bad scheduling. Patrick Donnelly
06:58 PM Bug #48839 (Resolved): qa: Error: Unable to find a match: cephfs-top
... Patrick Donnelly
06:40 PM Documentation #48838 (In Progress): document ms_mode options in mount.ceph manpage
Jeff Layton
06:37 PM Documentation #48838 (Resolved): document ms_mode options in mount.ceph manpage
We recently merged a patch to support the ms_mode= option in mount.ceph. Document the new option in the manpage. Jeff Layton
06:20 PM Backport #48837 (Resolved): nautilus: have mount helper pick appropriate mon sockets for ms_mode ...
https://github.com/ceph/ceph/pull/39133 Backport Bot
06:20 PM Backport #48836 (Resolved): octopus: have mount helper pick appropriate mon sockets for ms_mode v...
https://github.com/ceph/ceph/pull/40763 Backport Bot
06:18 PM Bug #48835 (New): qa: add ms_mode random choice to kclient tests
Building on https://github.com/ceph/ceph/pull/38788 (#48765): modify kernel_client.py to conditionally set ms_mode fo... Patrick Donnelly
06:15 PM Bug #48765 (Pending Backport): have mount helper pick appropriate mon sockets for ms_mode value
Patrick Donnelly
06:13 PM Bug #48661 (Resolved): mds: reserved can be set on feature set
Patrick Donnelly
06:11 PM Bug #48834 (Fix Under Review): qa: MDS_SLOW_METADATA_IO with osd thrasher
Patrick Donnelly
06:09 PM Bug #48834 (Resolved): qa: MDS_SLOW_METADATA_IO with osd thrasher
... Patrick Donnelly
06:02 PM Bug #48833 (New): snap_rm hang during osd thrashing
... Patrick Donnelly
05:47 PM Bug #48832: qa: fsstress w/ valgrind causes MDS to be blocklisted
Similar one: /ceph/teuthology-archive/pdonnell-2021-01-10_18:07:36-fs-wip-pdonnell-testing-20210110.050947-distro-bas... Patrick Donnelly
05:46 PM Bug #48832 (New): qa: fsstress w/ valgrind causes MDS to be blocklisted
... Patrick Donnelly
05:42 PM Bug #48831 (New): qa: ERROR: test_snapclient_cache
... Patrick Donnelly
05:40 PM Bug #48830 (Resolved): pacific: qa: :ERROR: test_idempotency
... Patrick Donnelly
05:37 PM Bug #48772: qa: pjd: not ok 9, 44, 80
Also k-stock:... Patrick Donnelly
02:50 PM Bug #48772 (Triaged): qa: pjd: not ok 9, 44, 80
Both tests are with the kernel's testing branch. Patrick Donnelly
05:11 PM Bug #44100: cephfs rsync kworker high load.
I also encountered this on Ubuntu 20.04 @Linux 5.4.0-60-generic #67-Ubuntu SMP Tue Jan 5 18:31:36 UTC 2021@
The clus...
Jonas Jelten
04:39 PM Feature #48602 (Resolved): `cephfs-top` frontend utility
Patrick Donnelly
04:30 PM Bug #42887 (Won't Fix): tasks.cephfs.test_volume_client.TestVolumeClient: test_21501 No such file...
Patrick Donnelly
04:30 PM Bug #42061 (Won't Fix): volume_client: AssertionError: 237 != 8
Patrick Donnelly
04:28 PM Bug #42724 (Won't Fix): pybind/mgr/volumes: confirm backwards-compatibility of ceph_volume_client.py
Patrick Donnelly
02:58 PM Bug #48760 (Triaged): qa: Test failure: test_acls (tasks.cephfs.test_acls.TestACLs)
Patrick Donnelly
02:56 PM Bug #48763 (Need More Info): mds memory leak
Can you set `debug mds = 10` during the event so we can get an idea what the MDS is doing during this time, assuming ... Patrick Donnelly
02:54 PM Bug #48766 (Triaged): qa: Test failure: test_evict_client (tasks.cephfs.test_volume_client.TestVo...
Patrick Donnelly
02:47 PM Bug #48773 (Triaged): qa: scrub does not complete
Patrick Donnelly
02:45 PM Bug #48778 (Need More Info): Setting quota triggered ops storms on meta pool
This doesn't sound like a bug we'd expect. I think you may have some clients executing these setxattr requests but I'... Patrick Donnelly
08:53 AM Bug #47787: mgr/nfs: exercise host-level HA of NFS-Ganesha by killing the process
Tim Serong wrote:
> How was your test cluster deployed (vstart/cstart/something else)? I'd like to see if I'm able ...
Varsha Rao
06:55 AM Bug #47787: mgr/nfs: exercise host-level HA of NFS-Ganesha by killing the process
I tried reproducing this (admittedly with a downstream SUSE build), and TERM seemed to have no effect, while KILL *di... Tim Serong
08:03 AM Feature #48404 (In Progress): client: add a ceph.caps vxattr
Kotresh Hiremath Ravishankar
08:02 AM Backport #48195: nautilus: mgr/volumes: allow/deny r/rw access of auth IDs to subvolume and subvo...
Patrick/Ramana,
I am planning to backport this along with [1] which is under review. The fix [1] persists the auth...
Kotresh Hiremath Ravishankar
08:02 AM Backport #48196: octopus: mgr/volumes: allow/deny r/rw access of auth IDs to subvolume and subvol...
Patrick/Ramana,
I am planning to backport this along with [1] which is under review. The fix [1] persists the auth...
Kotresh Hiremath Ravishankar

01/10/2021

05:04 AM Bug #47294: client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Adam Emerson wrote:
> This doesn't look quite the same? Before we were having the operation file with a definite tim...
Patrick Donnelly
04:47 AM Bug #47294: client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
This doesn't look quite the same? Before we were having the operation file with a definite timeout and I can't find t... Adam Emerson

01/09/2021

03:10 AM Bug #48811 (Closed): qa: fs/snaps/snaptest-realm-split.sh hang
https://github.com/ceph/ceph/pull/38732 Patrick Donnelly
03:07 AM Bug #48757 (Resolved): qa: "[WRN] Replacing daemon mds.d as rank 0 with standby daemon mds.f"
Patrick Donnelly

01/08/2021

11:20 PM Backport #48814 (Resolved): nautilus: mds: spurious wakeups in cache upkeep
https://github.com/ceph/ceph/pull/39130 Backport Bot
11:20 PM Backport #48813 (Resolved): octopus: mds: spurious wakeups in cache upkeep
https://github.com/ceph/ceph/pull/40743 Backport Bot
11:16 PM Bug #48753 (Pending Backport): mds: spurious wakeups in cache upkeep
Patrick Donnelly
11:14 PM Bug #48707 (Resolved): client: unmount() doesn't dump the cache
Patrick Donnelly
11:07 PM Bug #48812 (Resolved): qa: test_scrub_pause_and_resume_with_abort failure
... Patrick Donnelly
11:06 PM Bug #48365: qa: ffsb build failure on CentOS 8.2
Haven't seen this failure recently but did get this:... Patrick Donnelly
11:00 PM Bug #48811: qa: fs/snaps/snaptest-realm-split.sh hang
Probably related group of failures:... Patrick Donnelly
10:57 PM Bug #48811 (Closed): qa: fs/snaps/snaptest-realm-split.sh hang
... Patrick Donnelly
08:12 PM Bug #48808 (Fix Under Review): mon/MDSMonitor: `fs rm` is not idempotent
Patrick Donnelly
08:04 PM Bug #48808 (Resolved): mon/MDSMonitor: `fs rm` is not idempotent
... Patrick Donnelly
07:27 PM Bug #48756 (Resolved): qa: kclient does not synchronously write with O_DIRECT
Patrick Donnelly
07:26 PM Fix #48121 (Resolved): qa: merge fs/multimds suites
Patrick Donnelly
06:47 PM Bug #48702 (Resolved): qa: fwd_scrub should only scrub rank 0
Patrick Donnelly
06:39 PM Bug #48514 (Resolved): mgr/nfs: Don't prefix 'ganesha-' to cluster id
Patrick Donnelly
06:37 PM Bug #47294 (Need More Info): client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Looks like this is still here, maybe more racy than before:... Patrick Donnelly
06:26 PM Bug #48805 (Resolved): mds: "cluster [WRN] Scrub error on inode 0x1000000039d (/client.0/tmp/blog...
... Patrick Donnelly
06:19 PM Feature #18513 (Resolved): MDS: scrub: forward scrub reports missing backtraces on new files as d...
This is fixed in the current code. Patrick Donnelly
04:38 PM Fix #48802 (Resolved): mds: define CephFS errors that replace standard errno values
CephFS protocol depends on the errno numbers which may vary by operating system. Copy the Linux values we use into in... Patrick Donnelly

01/07/2021

09:58 PM Feature #48791: mds: support file block size
Technically, I think the blocksize can be anything >= 8 bytes or so. Too small or large a block will be cumbersome to... Jeff Layton
09:53 PM Feature #48791 (Rejected): mds: support file block size
The new fscrypt feature in the kernel client that is under development needs to be able to prevent the MDS from trunc... Patrick Donnelly
05:57 PM Bug #48203 (Resolved): qa: quota failure
Luis Henriques wrote:
> Patrick Donnelly wrote:
> > Hey Luis, I think this is still broken; the revert didn't work:...
Patrick Donnelly
04:42 PM Bug #48203: qa: quota failure
Patrick Donnelly wrote:
> Hey Luis, I think this is still broken; the revert didn't work: https://pulpito.ceph.com/t...
Luis Henriques
03:50 PM Bug #48203 (Need More Info): qa: quota failure
Hey Luis, I think this is still broken; the revert didn't work: https://pulpito.ceph.com/teuthology-2021-01-03_03:15:... Patrick Donnelly
12:22 PM Backport #48457 (Resolved): nautilus: client: fix crash when doing remount in none fuse case
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/38467
m...
Nathan Cutler
12:22 PM Backport #48110 (Resolved): nautilus: client: ::_read fails to advance pos at EOF checking
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37991
m...
Nathan Cutler
12:21 PM Backport #48097 (Resolved): nautilus: osdc: FAILED ceph_assert(bh->waitfor_read.empty())
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37988
m...
Nathan Cutler
12:21 PM Backport #48095 (Resolved): nautilus: mds: fix file recovery crash after replaying delayed requests
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37986
m...
Nathan Cutler
05:28 AM Bug #48778 (Need More Info): Setting quota triggered ops storms on meta pool
A few days ago, I toyed with setting a quota on a handful (less than 60) directories on my CephFS filesystem. At the ... Gerry D

01/06/2021

11:20 PM Backport #48457: nautilus: client: fix crash when doing remount in none fuse case
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/38467
merged
Yuri Weinstein
11:20 PM Backport #48110: nautilus: client: ::_read fails to advance pos at EOF checking
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37991
merged
Yuri Weinstein
11:19 PM Backport #48097: nautilus: osdc: FAILED ceph_assert(bh->waitfor_read.empty())
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37988
merged
Yuri Weinstein
11:19 PM Backport #48095: nautilus: mds: fix file recovery crash after replaying delayed requests
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37986
merged
Yuri Weinstein
09:35 PM Documentation #48585 (Resolved): mds_cache_trim_decay_rate misnamed?
Patrick Donnelly
08:31 PM Bug #48773 (In Progress): qa: scrub does not complete
... Patrick Donnelly
08:29 PM Bug #48772 (Need More Info): qa: pjd: not ok 9, 44, 80
... Patrick Donnelly
08:25 PM Bug #48771 (New): qa: iogen: workload fails to cause balancing
Not really a bug but it causes a test failure and is worthy of investigation:... Patrick Donnelly
07:54 PM Bug #48765 (Fix Under Review): have mount helper pick appropriate mon sockets for ms_mode value
Jeff Layton
02:13 PM Bug #48765 (Resolved): have mount helper pick appropriate mon sockets for ms_mode value
Ilya recently added msgr2 support to the kclient, but the mount helper still ignores any v2 addresses when mounting. ... Jeff Layton
07:12 PM Bug #48770 (Fix Under Review): qa: "Test failure: test_hole (tasks.cephfs.test_failover.TestClust...
Patrick Donnelly
06:56 PM Bug #48770 (Resolved): qa: "Test failure: test_hole (tasks.cephfs.test_failover.TestClusterResize)"
... Patrick Donnelly
05:50 PM Bug #48766: qa: Test failure: test_evict_client (tasks.cephfs.test_volume_client.TestVolumeClient)
master failure: /ceph/teuthology-archive/teuthology-2020-12-27_03:15:03-fs-master-distro-basic-smithi/5738903/teuthol... Patrick Donnelly
02:23 PM Bug #48766 (Duplicate): qa: Test failure: test_evict_client (tasks.cephfs.test_volume_client.Test...
test_evict_client fails in [1] and [2].
[1] http://qa-proxy.ceph.com/teuthology/jcollin-2021-01-05_16:12:23-fs-wip...
Jos Collin
03:02 PM Bug #48760: qa: Test failure: test_acls (tasks.cephfs.test_acls.TestACLs)
failure in master:
/ceph/teuthology-archive/teuthology-2021-01-05_03:15:02-fs-master-distro-basic-smithi/5754681/t...
Patrick Donnelly
04:37 AM Bug #48760 (Can't reproduce): qa: Test failure: test_acls (tasks.cephfs.test_acls.TestACLs)
Job 5756405 [1] fails with below error.
[1] http://qa-proxy.ceph.com/teuthology/jcollin-2021-01-05_16:12:23-fs-wip...
Jos Collin
02:51 PM Documentation #48531 (Resolved): doc/cephfs: "ceph fs new" command is, ironically, old. The new (...
Patrick Donnelly
02:20 PM Feature #44928 (Fix Under Review): mgr/volumes: evict clients based on auth ID and subvolume mounted
Kotresh Hiremath Ravishankar
01:51 PM Bug #45344: doc: Table Of Contents doesn't work
I spoke to Patrick (the creator and owner of the CephFS documentation) about this, and for the time being, the rst fi... Zac Dover
01:43 PM Bug #48763: mds memory leak
... Anonymous
09:47 AM Bug #48763 (Need More Info): mds memory leak
I have a possible memory leak in 14.2.10 mds. MDS suddenly uses 107GB Ram, before around 70.
Once mds starts its eat...
Anonymous
12:57 PM Bug #44565 (In Progress): src/mds/SimpleLock.h: 528: FAILED ceph_assert(state == LOCK_XLOCK || st...
Jan Fajerski

01/05/2021

09:45 PM Bug #48756 (Fix Under Review): qa: kclient does not synchronously write with O_DIRECT
Patrick Donnelly
08:15 PM Bug #48756: qa: kclient does not synchronously write with O_DIRECT
Trying to reproduce on master: https://pulpito.ceph.com/pdonnell-2021-01-05_20:15:07-fs:workload-master-distro-basic-... Patrick Donnelly
08:09 PM Bug #48756 (Resolved): qa: kclient does not synchronously write with O_DIRECT
... Patrick Donnelly
08:35 PM Bug #48757 (Fix Under Review): qa: "[WRN] Replacing daemon mds.d as rank 0 with standby daemon md...
Patrick Donnelly
08:30 PM Bug #48757 (Resolved): qa: "[WRN] Replacing daemon mds.d as rank 0 with standby daemon mds.f"
/ceph/teuthology-archive/pdonnell-2020-12-24_22:49:03-fs:workload-wip-pdonnell-testing-20201224.195406-distro-basic-s... Patrick Donnelly
05:07 PM Bug #48753 (Fix Under Review): mds: spurious wakeups in cache upkeep
Patrick Donnelly
05:06 PM Bug #48753 (Resolved): mds: spurious wakeups in cache upkeep
... Patrick Donnelly
04:39 PM Documentation #48731 (Resolved): mgr/nfs: Add info related to rook, clarify pseudo path and dashb...
Patrick Donnelly
03:02 PM Feature #46074 (Resolved): mds: provide altrenatives to increase the total cephfs subvolume snaps...
While running with --resolve-parent, the script "backport-create-issue" noticed that all backports of this issue are ... Nathan Cutler
03:00 PM Bug #48555 (Resolved): pybind/ceph_volume_client: allows authorize on auth_ids not created throug...
While running with --resolve-parent, the script "backport-create-issue" noticed that all backports of this issue are ... Nathan Cutler
02:08 PM Documentation #48531 (Fix Under Review): doc/cephfs: "ceph fs new" command is, ironically, old. T...
Zac Dover

01/04/2021

02:44 PM Bug #48673 (Need More Info): High memory usage on standby replay MDS
Patrick Donnelly
02:44 PM Bug #48711 (Triaged): mds: standby-replay mds abort when replay metablob
Patrick Donnelly
11:46 AM Feature #45746 (In Progress): mgr/nfs: Add interface to update export
Varsha Rao
10:09 AM Feature #48736 (Fix Under Review): qa: enable debug loglevel kclient test suits
Xiubo Li
04:24 AM Feature #48736 (In Progress): qa: enable debug loglevel kclient test suits
Xiubo Li
03:43 AM Feature #48736 (Resolved): qa: enable debug loglevel kclient test suits
This is helpful when debugging and resolving bugs. Xiubo Li

12/31/2020

12:25 PM Documentation #48731 (In Progress): mgr/nfs: Add info related to rook, clarify pseudo path and da...
Varsha Rao
12:19 PM Documentation #48731 (Resolved): mgr/nfs: Add info related to rook, clarify pseudo path and dashb...
Varsha Rao

12/30/2020

01:46 AM Bug #48679 (Fix Under Review): client: items pinned in cache preventing unmount
Xiubo Li
01:35 AM Bug #48679: client: items pinned in cache preventing unmount
Xiubo Li wrote:
> For example for the inode 0x10000000e51:
>
> [...]
>
> Because it has the Fb cap, so the flu...
Xiubo Li

12/28/2020

09:27 AM Backport #47085 (Resolved): octopus: common: validate type CephBool cause 'invalid command json'
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37362
m...
Nathan Cutler
09:26 AM Backport #47095 (Resolved): octopus: mds: provide altrenatives to increase the total cephfs subvo...
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/38553
m...
Nathan Cutler
09:25 AM Backport #48372 (Resolved): octopus: client: dump which fs is used by client for multiple-fs
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/38551
m...
Nathan Cutler

12/24/2020

11:27 AM Bug #48679: client: items pinned in cache preventing unmount
For example for the inode 0x10000000e51:... Xiubo Li
04:35 AM Bug #47662 (Resolved): mds: try to replicate hot dir to restarted MDS
Patrick Donnelly
04:33 AM Fix #48053 (Resolved): qa: update test_readahead to work with the kernel
Patrick Donnelly
04:32 AM Bug #48701 (Resolved): pybind/cephfs: MCommand message is constructed with command separated into...
Patrick Donnelly
04:20 AM Bug #47294 (Resolved): client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Patrick Donnelly
02:55 AM Bug #48711 (Closed): mds: standby-replay mds abort when replay metablob
Ceph Version 14.2.15
OS: CentOS 7.6.1810
We create a fs that have three active mds, three standby-replay mds, three...
haitao chen
12:24 AM Feature #44192 (Fix Under Review): mds: stable multimds scrub
Patrick Donnelly

12/23/2020

08:25 AM Bug #48707 (Fix Under Review): client: unmount() doesn't dump the cache
Xiubo Li
08:23 AM Bug #48707 (Resolved): client: unmount() doesn't dump the cache
The delay_put_inodes() will be called by the tick() periodically
per second, and when the _unmount() is waiting for ...
Xiubo Li
07:53 AM Bug #48706: mgr/nfs: Does not detect exports created by dashboard
Dashboard can detect volume/nfs exports Varsha Rao
07:50 AM Bug #48706 (New): mgr/nfs: Does not detect exports created by dashboard
... Varsha Rao
06:28 AM Bug #48679 (In Progress): client: items pinned in cache preventing unmount
Xiubo Li
05:54 AM Bug #48559 (Fix Under Review): qa: ERROR: test_damaged_dentry, KeyError: 'passed_validation'
Xiubo Li
02:40 AM Bug #44565: src/mds/SimpleLock.h: 528: FAILED ceph_assert(state == LOCK_XLOCK || state == LOCK_XL...
We have just hit this in 14.2.11, with 3 active mds.
mds log is at ceph-post-file: b1a56b74-6fbe-41bb-adcd-183695c39...
Dan van der Ster

12/22/2020

09:14 PM Feature #48704 (New): mds: recall caps proportional to the number issued
mds_recall_max_caps may wipe out the client cache for small clients. It may also not be large enough for very aggress... Patrick Donnelly
06:30 PM Backport #48703 (Rejected): octopus: mgr/nfs: Add tests for readonly exports
Backport Bot
06:26 PM Feature #48622 (Pending Backport): mgr/nfs: Add tests for readonly exports
Patrick Donnelly
06:01 PM Bug #48702 (Fix Under Review): qa: fwd_scrub should only scrub rank 0
Patrick Donnelly
05:42 PM Bug #48702 (Resolved): qa: fwd_scrub should only scrub rank 0
... Patrick Donnelly
05:45 PM Backport #47085: octopus: common: validate type CephBool cause 'invalid command json'
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37362
merged
Yuri Weinstein
05:30 PM Bug #48701 (Fix Under Review): pybind/cephfs: MCommand message is constructed with command separa...
Patrick Donnelly
05:28 PM Bug #48701 (Resolved): pybind/cephfs: MCommand message is constructed with command separated into...
... Patrick Donnelly
04:47 PM Backport #48111 (Resolved): octopus: doc: document MDS recall configurations
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/38202
m...
Nathan Cutler
04:44 PM Backport #48111: octopus: doc: document MDS recall configurations
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/38202
merged
Yuri Weinstein
04:47 PM Backport #47095: octopus: mds: provide altrenatives to increase the total cephfs subvolume snapsh...
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/38553
merged
Yuri Weinstein
04:47 PM Backport #48191 (Resolved): octopus: mds: throttle workloads which acquire caps faster than the c...
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/38095
m...
Nathan Cutler
04:44 PM Backport #48191: octopus: mds: throttle workloads which acquire caps faster than the client can r...
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/38095
merged
Yuri Weinstein
04:47 PM Backport #48109 (Resolved): octopus: client: ::_read fails to advance pos at EOF checking
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37989
m...
Nathan Cutler
04:43 PM Backport #48109: octopus: client: ::_read fails to advance pos at EOF checking
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37989
merged
Yuri Weinstein
04:47 PM Backport #48098 (Resolved): octopus: osdc: FAILED ceph_assert(bh->waitfor_read.empty())
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37987
m...
Nathan Cutler
04:43 PM Backport #48098: octopus: osdc: FAILED ceph_assert(bh->waitfor_read.empty())
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37987
merged
Yuri Weinstein
04:46 PM Backport #48372: octopus: client: dump which fs is used by client for multiple-fs
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/38551
merged
Yuri Weinstein
04:46 PM Backport #48096 (Resolved): octopus: mds: fix file recovery crash after replaying delayed requests
This update was made using the script "backport-resolve-issue".
backport PR https://github.com/ceph/ceph/pull/37985
m...
Nathan Cutler
04:42 PM Backport #48096: octopus: mds: fix file recovery crash after replaying delayed requests
Nathan Cutler wrote:
> https://github.com/ceph/ceph/pull/37985
merged
Yuri Weinstein
04:45 PM Bug #48524: octopus: run_shell() got an unexpected keyword argument 'timeout'
https://github.com/ceph/ceph/pull/38550 merged Yuri Weinstein
02:33 PM Feature #44928 (In Progress): mgr/volumes: evict clients based on auth ID and subvolume mounted
Kotresh Hiremath Ravishankar
01:46 PM Bug #48700 (Closed): client: Client::rmdir() may fail to remove a snapshot
Call to Client::may_delete() from Client::rmdir() is done here https://github.com/ceph/ceph/blob/master/src/client/Cl... Venky Shankar

12/21/2020

04:26 PM Bug #48673: High memory usage on standby replay MDS
Thanks for the information. There were a few fixes in v15.2.8 relating to memory consumption for the MDS which may be... Patrick Donnelly
04:17 PM Bug #48673: High memory usage on standby replay MDS
Patrick Donnelly wrote:
> Please share `ceph versions` and `ceph fs dump`.
>
> I believe we've recently fixed som...
Daniel Persson
02:53 PM Bug #48673: High memory usage on standby replay MDS
Daniel Persson wrote:
> Hi.
>
> We have recently installed a Ceph cluster and with about 27M objects. The filesys...
Patrick Donnelly
04:12 PM Fix #48121 (Fix Under Review): qa: merge fs/multimds suites
Patrick Donnelly
02:52 PM Bug #48679: client: items pinned in cache preventing unmount
Patrick, this one seems similiar to the one I have fixed before, I will take it.
Thanks.
Xiubo Li
02:41 PM Bug #48679: client: items pinned in cache preventing unmount
... Patrick Donnelly
02:46 PM Feature #48619 (In Progress): client: track (and forward to MDS) average read/write/metadata latency
Patrick Donnelly
06:38 AM Bug #47294 (Fix Under Review): client: thread hang in Client::_setxattr_maybe_wait_for_osdmap
Fix in https://github.com/ceph/ceph/pull/38668 Adam Emerson
03:46 AM Bug #48559: qa: ERROR: test_damaged_dentry, KeyError: 'passed_validation'
The commit(03908aa04344) has removed the code which was dumpping the scrub detail result. And currently only the foll... Xiubo Li

12/19/2020

09:26 PM Feature #7320 (Fix Under Review): qa: thrash directory fragmentation
Patrick Donnelly
06:36 PM Fix #48683 (Resolved): mds/MDSMap: print each flag value in MDSMap::dump
Don't require operators to do bitwise arithmetic on the "flags" field. Print each flag.
https://github.com/ceph/ce...
Patrick Donnelly
06:35 PM Feature #48682 (Resolved): MDSMonitor: add command to print fs flags
From this list:
https://github.com/ceph/ceph/blob/master/src/include/ceph_fs.h#L275-L285
Patrick Donnelly

12/18/2020

09:27 PM Bug #48517 (Resolved): mds: "CDir.cc: 1530: FAILED ceph_assert(!is_complete())"
Patrick Donnelly
09:25 PM Feature #22477 (Resolved): multifs: remove multifs experimental warnings
Patrick Donnelly
09:15 PM Bug #48680 (New): mds: scrubbing stuck "scrub active (0 inodes in the stack)"
... Patrick Donnelly
09:10 PM Bug #48679 (Resolved): client: items pinned in cache preventing unmount
... Patrick Donnelly
09:06 PM Bug #48678 (In Progress): client: spins on tick interval
... Patrick Donnelly
02:41 PM Bug #48501 (Fix Under Review): pybind/mgr/volumes: inherited snapshots should be filtered out of ...
Kotresh Hiremath Ravishankar
08:21 AM Bug #48673 (Pending Backport): High memory usage on standby replay MDS
Hi.
We have recently installed a Ceph cluster and with about 27M objects. The filesystem seems to have 15M files.
...
Daniel Persson
04:33 AM Feature #44931 (Fix Under Review): mgr/volumes: get the list of auth IDs that have been granted a...
Kotresh Hiremath Ravishankar

12/17/2020

11:39 PM Bug #21539: man: missing man page for mount.fuse.ceph
Adding this to the packaging in https://github.com/ceph/ceph/pull/38642 Ken Dreyer
07:29 PM Bug #48661 (Fix Under Review): mds: reserved can be set on feature set
Patrick Donnelly
07:28 PM Bug #48661 (Resolved): mds: reserved can be set on feature set
... Patrick Donnelly
05:40 PM Backport #48638 (Resolved): nautilus: pybind/ceph_volume_client: allows authorize on auth_ids not...
Nathan Cutler
05:39 PM Backport #48638 (In Progress): nautilus: pybind/ceph_volume_client: allows authorize on auth_ids ...
Nathan Cutler
12:06 AM Backport #48638 (Resolved): nautilus: pybind/ceph_volume_client: allows authorize on auth_ids not...
Patrick Donnelly
12:05 AM Backport #48638 (Resolved): nautilus: pybind/ceph_volume_client: allows authorize on auth_ids not...
... Backport Bot
04:15 AM Backport #48644 (Resolved): octopus: client: ceph.dir.entries does not acquire necessary caps
https://github.com/ceph/ceph/pull/38949 Backport Bot
04:15 AM Backport #48643 (Resolved): nautilus: client: ceph.dir.entries does not acquire necessary caps
https://github.com/ceph/ceph/pull/38950 Backport Bot
04:12 AM Bug #48313 (Pending Backport): client: ceph.dir.entries does not acquire necessary caps
Patrick Donnelly
04:11 AM Feature #17856 (Resolved): qa: background cephfs forward scrub teuthology task
Patrick Donnelly
04:10 AM Backport #48642 (Resolved): octopus: Client: the directory's capacity will not be updated after w...
https://github.com/ceph/ceph/pull/38947 Backport Bot
04:10 AM Backport #48641 (Resolved): nautilus: Client: the directory's capacity will not be updated after ...
https://github.com/ceph/ceph/pull/38948 Backport Bot
04:09 AM Bug #48318 (Pending Backport): Client: the directory's capacity will not be updated after write d...
Patrick Donnelly
02:19 AM Backport #48639 (Resolved): luminous: pybind/ceph_volume_client: allows authorize on auth_ids not...
Patrick Donnelly
12:10 AM Backport #48639 (Resolved): luminous: pybind/ceph_volume_client: allows authorize on auth_ids not...
... Backport Bot
12:06 AM Backport #48637 (Resolved): octopus: pybind/ceph_volume_client: allows authorize on auth_ids not ...
Patrick Donnelly
12:05 AM Backport #48637 (Resolved): octopus: pybind/ceph_volume_client: allows authorize on auth_ids not ...
... Backport Bot
12:05 AM Bug #48555: pybind/ceph_volume_client: allows authorize on auth_ids not created through ceph_volu...
... Patrick Donnelly
12:03 AM Bug #48555 (Pending Backport): pybind/ceph_volume_client: allows authorize on auth_ids not create...
Patrick Donnelly
12:03 AM Bug #48555 (Resolved): pybind/ceph_volume_client: allows authorize on auth_ids not created throug...
Backports done manually. Patrick Donnelly

12/16/2020

10:19 PM Backport #48634 (In Progress): nautilus: qa: tox failures
Patrick Donnelly
10:15 PM Backport #48634 (Resolved): nautilus: qa: tox failures
https://github.com/ceph/ceph/pull/38627 Backport Bot
10:18 PM Backport #48635 (In Progress): octopus: qa: tox failures
Patrick Donnelly
10:15 PM Backport #48635 (Resolved): octopus: qa: tox failures
https://github.com/ceph/ceph/pull/38626 Backport Bot
10:14 PM Bug #48633 (Pending Backport): qa: tox failures
Neha Ojha
08:44 PM Bug #48633 (Fix Under Review): qa: tox failures
Patrick Donnelly
08:43 PM Bug #48633 (Resolved): qa: tox failures
... Patrick Donnelly
02:52 PM Backport #47158 (In Progress): octopus: mgr/volumes: Mark subvolumes with ceph.dir.subvolume vxat...
Shyamsundar Ranganathan
01:50 PM Feature #48622 (Fix Under Review): mgr/nfs: Add tests for readonly exports
Varsha Rao
08:23 AM Feature #48622 (Resolved): mgr/nfs: Add tests for readonly exports
Varsha Rao
09:50 AM Bug #47787: mgr/nfs: exercise host-level HA of NFS-Ganesha by killing the process
I can reproduce it with SIGTERM... Varsha Rao
06:52 AM Feature #48619: client: track (and forward to MDS) average read/write/metadata latency
Xiubo suggested that the client also sends min/max and stddev. Venky Shankar
05:17 AM Feature #48619 (Pending Backport): client: track (and forward to MDS) average read/write/metadata...
Client already tracks cumulative read/write/metadata latencies. However, average latencies are much more useful to th... Venky Shankar
06:01 AM Tasks #48620 (In Progress): mds: break the mds_lock or get rid of the mds_lock for some code
Xiubo Li
05:48 AM Tasks #48620 (In Progress): mds: break the mds_lock or get rid of the mds_lock for some code
Xiubo Li
05:54 AM Bug #48559 (In Progress): qa: ERROR: test_damaged_dentry, KeyError: 'passed_validation'
Xiubo Li
03:18 AM Bug #48517 (Fix Under Review): mds: "CDir.cc: 1530: FAILED ceph_assert(!is_complete())"
Xiubo Li

12/15/2020

01:02 PM Feature #48602 (Resolved): `cephfs-top` frontend utility
The plumbing work for tracking (client) metrics in the MDS is already done and mgr/stats module provides an interface... Venky Shankar
12:41 PM Documentation #48585 (Fix Under Review): mds_cache_trim_decay_rate misnamed?
No other places, just being more explicit would be helpful I think. Jan Fajerski
 

Also available in: Atom