Activity
From 09/15/2013 to 10/14/2013
10/14/2013
- 08:24 PM Bug #5025: samba smbtorture lock test fails on kclient
- in commit:f5685013a4ff4f569413945f380dd48b6cdbfaad
- 07:01 PM Bug #5025 (Resolved): samba smbtorture lock test fails on kclient
- 09:41 AM Bug #6396 (Resolved): mds: recovery hits assert(!segment.empty()) when reissuing caps
- Backported to dumpling as commit:cd1c3c9e00e90b19e83c1f11a48e516a7de93665
- 09:27 AM Bug #1117: mds: rename rollback broken on slaves during replay
- Which patch[es]?
- 07:29 AM Bug #1117 (Resolved): mds: rename rollback broken on slaves during replay
- 09:25 AM Bug #3738: kclient fsx truncate/write multi-client race
- Is this a side effect of 3c3b2ceb03e7294704f5bf3e1e420012a0166585, or some other patch? Please reference them when cl...
- 07:24 AM Bug #3738 (Resolved): kclient fsx truncate/write multi-client race
- now mds revokes Fw before doing truncate
- 07:40 AM Bug #2385 (Can't reproduce): max mds = 2, mds hang and crash
- 07:34 AM Bug #5036 (Resolved): `ls` hangs on random folder
- 07:34 AM Bug #2019 (Resolved): mds: CInode::filelock stuck in sync->mix
- 07:31 AM Bug #3088 (Resolved): NULL pointer dereference at ceph_d_prune
- 07:28 AM Bug #429 (Resolved): mds: fix rstat propogation into past parents
- haven't seen any rstat error for months
- 07:25 AM Bug #3681 (Resolved): kclient fsx fails nightly
10/11/2013
- 09:10 AM Feature #6511 (Rejected): MDS: add special purging options for testing
- We do a lot of stuff async with our journal, stray inodes, etc, that we need a good way to test. Let this ticket serv...
- 07:33 AM Bug #4405 (Pending Backport): MDCache::populate_mydir can loop forever
10/10/2013
- 10:42 PM Bug #6396 (Pending Backport): mds: recovery hits assert(!segment.empty()) when reissuing caps
- This seems to be showing up in our nightlies a fair bit still; it should be backported.
- 01:18 PM Bug #4405 (Fix Under Review): MDCache::populate_mydir can loop forever
- He's got a patch in wip-4405 that he thinks would solve the loading problem; I think he might be right. The other iss...
- 11:05 AM Bug #6460: ceph-fuse: xlist crash in ~Inode on osd_op_reply
- In general I'm not usually too worried about FS backports since anybody using the FS should be keeping up to date wit...
- 04:27 AM Bug #6460 (Resolved): ceph-fuse: xlist crash in ~Inode on osd_op_reply
10/09/2013
- 08:16 PM Bug #6460 (Pending Backport): ceph-fuse: xlist crash in ~Inode on osd_op_reply
- Probably already fixed by https://github.com/ceph/ceph/pull/590. Do we need to backport MDS fixes to old version?
- 02:02 PM Bug #4832: mds: failed auth_unpin assert
- I've backported these two patches to the cuttlefish branch as well.
(Plus a cuttlefish-4832 branch on top of v0.61.8...
10/04/2013
- 08:41 AM Bug #5290 (Can't reproduce): mds: crash whilst trying to reconnect
- 07:24 AM Bug #5290: mds: crash whilst trying to reconnect
- I certainly haven't been hit by this again, so if you consider it resolved...
10/03/2013
- 07:50 PM Bug #5250: ceph-mds 0.61.2 aborts on start
- I just updated to 0.67.3 and I want to confirm that I still have to recompile to get around the abort.
0> 201... - 06:11 PM Bug #6473: multimds + ceph-fuse: fsstress gets ENOTEMPTY on final rm -r
- ubuntu@teuthology:/a/sage-2013-10-03_17:25:57-marginal:multimds-master-testing-basic-/32499
- 06:11 PM Bug #6473 (Can't reproduce): multimds + ceph-fuse: fsstress gets ENOTEMPTY on final rm -r
- ...
- 01:34 PM Bug #5025: samba smbtorture lock test fails on kclient
- http://qa-proxy.ceph.com/teuthology/teuthology-2013-10-02_23:01:08-fs-master-testing-basic-plana/31442/
- 01:33 PM Bug #5025: samba smbtorture lock test fails on kclient
- oops, still fails on kclient.
- 01:33 PM Bug #2825 (Resolved): File lock doesn't work properly
10/02/2013
- 10:21 PM Bug #5753: ceph-fuse: segfault when getting back a traceless rename op
- and /a/teuthology-2013-10-01_23:01:09-fs-next-testing-basic-plana/29640
- 01:00 PM Bug #5753: ceph-fuse: segfault when getting back a traceless rename op
- Showed up again, /a/teuthology-2013-09-27_19:01:27-fs-dumpling-testing-basic-plana/21601
- 09:53 PM Bug #6458 (New): journaler: journal too short during replay
- Wow, that explanation of what was going on was so very wrong. Now I'm just not sure how this could have occurred.
- 06:27 PM Bug #6458: journaler: journal too short during replay
- Made a pull request too: https://github.com/ceph/ceph/pull/683
- 06:26 PM Bug #6458: journaler: journal too short during replay
- Yep, it ran pjd successfully; plenty of journal commits there!
- 06:19 PM Bug #6458 (Fix Under Review): journaler: journal too short during replay
- Pushed a patch to wip-journaler-safety, commit:a0ba5c66162af720627fcf7ba63fdc76ac97f568. I'm setting up a basic funct...
- 05:58 PM Bug #6458 (In Progress): journaler: journal too short during replay
- This is a bit more complicated than we described — we do not in fact blindly write the write_pos to our head object; ...
- 11:19 AM Bug #6458 (New): journaler: journal too short during replay
- Urgh, that last comment was mistaken.
- 10:56 AM Bug #6458 (Rejected): journaler: journal too short during replay
- That is not what happened; the underlying objects were inconsistent in RADOS.
- 10:07 AM Bug #6458 (Can't reproduce): journaler: journal too short during replay
- Got a report on irc from a user whose log was 611 bytes shorter than the header indicated it should be. His guess was...
- 02:44 PM Bug #1596 (Can't reproduce): mds crash during ffsb on kernel client in CInode::is_frozen
- 02:44 PM Bug #1601 (Can't reproduce): mds crash during snaps workunit
- 02:44 PM Bug #1752 (Can't reproduce): ceph-fuse isn't releasing caps without flushing data?
- 02:43 PM Bug #3601: client: With multiple clients, file remove doesn't free up space
- repushing to wip-fuse for testing
- 02:38 PM Bug #5025 (Resolved): samba smbtorture lock test fails on kclient
- lock test was fixed a while ago, commit:476e4902907dfadb3709ba820453299ececf990b test is reenabled in the suite.
- 12:54 PM Bug #6460 (Resolved): ceph-fuse: xlist crash in ~Inode on osd_op_reply
- ...
- 12:18 PM Bug #3681: kclient fsx fails nightly
- added fsx back into the kcephfs test suite. reportedly fsx now passes, but we should verify before closing this bug.
- 12:13 PM Bug #5037 (Can't reproduce): Ceph-MDS asserts after upgrade 0.56.2 -> 0.56.6
- 12:12 PM Bug #5162 (Can't reproduce): File is locked unexpected and not released anymore
- 12:07 PM Bug #5033 (Can't reproduce): oops in ceph_put_wrbuffer_cap_refs
- 12:07 PM Bug #5290: mds: crash whilst trying to reconnect
- i'm inclined to call this can't reproduce. there was a locking fix recently that covered the session_map too that co...
- 12:05 PM Bug #5418 (Resolved): kceph: crash in remove_session_caps
- 11:44 AM Feature #6332 (Resolved): mds: add config option disabling snapshots by default
- This got merged to master; I'm going to call that done since we're releasing Emperor soon-ish.
- 02:34 AM Bug #6450: Kernel bugs in 3.12-rc1, taking 2 hosts (and 1 following) down
- we have now upgraded the complete cluster to Dumpling (0.67.3) (also due to other problems we have experienced in the...
10/01/2013
- 09:11 PM Bug #6396 (Resolved): mds: recovery hits assert(!segment.empty()) when reissuing caps
- yeah looks good!
- 09:07 PM Bug #6396: mds: recovery hits assert(!segment.empty()) when reissuing caps
- 09:07 PM Bug #6396: mds: recovery hits assert(!segment.empty()) when reissuing caps
- By reading the mds log for #5458, I think these issues should be mostly fixed by commit b144170544(mds: properly retu...
- 09:09 PM Bug #6349 (Duplicate): MDS: failed assert !segments.empty() while rejoining after being standby-r...
- dup #6396
- 09:08 PM Bug #5458 (Duplicate): mds: standby-replay -> replay takeover does not handle racing expire/trim
- dup #6396
- 06:47 AM Bug #6450: Kernel bugs in 3.12-rc1, taking 2 hosts (and 1 following) down
- Jens-Christian Fischer wrote:
> Can I run a 0.67.3 MDS with the rest of the infrastructure on 0.61.8?
yes, you ca... - 05:34 AM Bug #6450: Kernel bugs in 3.12-rc1, taking 2 hosts (and 1 following) down
- Can I run a 0.67.3 MDS with the rest of the infrastructure on 0.61.8?
We are using the rc1 kernels in order to run... - 05:28 AM Bug #6450 (Need More Info): Kernel bugs in 3.12-rc1, taking 2 hosts (and 1 following) down
- The first warning was caused by a MDS bug. (you can try upgrading MDS 0.67.3 ) The rest BUGs did not look like ceph r...
- 04:57 AM Bug #6450 (Closed): Kernel bugs in 3.12-rc1, taking 2 hosts (and 1 following) down
- We are running 10 hosts with 74 OSDs on Ubuntu 13.04, Ceph 0.61.8 and Kernel 3.12-rc1
root@h5:~# ceph --version
...
09/26/2013
- 10:31 AM Bug #6399 (Resolved): snaptest-0.sh fails
- In master, commit:3cffff2a1abd266abbbc97effd73b79695d29682
- 09:34 AM Bug #6399 (Fix Under Review): snaptest-0.sh fails
- 08:27 AM Bug #6396: mds: recovery hits assert(!segment.empty()) when reissuing caps
- /a/teuthology-2013-09-20_23:01:15-fs-master-testing-basic-plana/8560
All these archive dirs we reference here are ...
09/25/2013
- 11:43 PM Bug #6396: mds: recovery hits assert(!segment.empty()) when reissuing caps
- where to get the mds log? (I can access your lab now)
- 01:06 PM Bug #6396 (Resolved): mds: recovery hits assert(!segment.empty()) when reissuing caps
- ...
- 04:53 PM Bug #6399 (Resolved): snaptest-0.sh fails
- MDS backtrace:...
- 12:53 PM Bug #6394 (Resolved): teuthology: bad dereference in mds thrasher
- ...
09/23/2013
- 09:25 PM Feature #6373 (In Progress): kcephfs: qa: test fscache
- 03:42 PM Bug #4221 (Resolved): MDS: LogEvent::decode needs to respect mds_log_skip_corrupt_events for DECO...
- Master commit:73289b34b0be5b6612e38944794d59b5e789f841
Dumpling commit:b3d3b3747c1eef695138dac828e5fcb435309c7b
09/20/2013
- 10:23 AM Feature #4354: mds: add an equivalent to the OSD OpTracker
- Got the OpTracker work rebased on master and pushed for review: wip-optracker, https://github.com/ceph/ceph/pull/610....
- 08:47 AM Bug #6284 (Resolved): client: failed xlist assert on put_inode()
09/19/2013
- 01:37 PM Feature #6332 (Fix Under Review): mds: add config option disabling snapshots by default
- wip-6332; https://github.com/ceph/ceph/pull/606
09/18/2013
- 02:38 PM Bug #6349 (Duplicate): MDS: failed assert !segments.empty() while rejoining after being standby-r...
- ...
09/17/2013
- 09:56 AM Feature #6332 (Resolved): mds: add config option disabling snapshots by default
- Right now snapshots are even more broken than they used to be (see #5250). Add a config option that controls whether ...
Also available in: Atom