Project

General

Profile

Activity

From 01/06/2016 to 02/04/2016

02/04/2016

11:56 PM Bug #14641: don't let users specify 0 on stripe count or object size
This is based on http://www.spinics.net/lists/ceph-users/msg25363.html
Perhaps the cluster is old and they broke i...
Greg Farnum
11:43 PM Bug #14641: don't let users specify 0 on stripe count or object size
We already validate layouts in Server.cc. We *could* also validate them on load from disk or during scrub...? John Spray
03:15 PM Bug #14641 (Duplicate): don't let users specify 0 on stripe count or object size
Those are fairly nonsensical. Amongst other things, they induce a divide-by-zero in the StrayManager right now. Greg Farnum
03:16 PM Feature #14642 (New): Validate layouts everywhere we load them
See _calculate_ops_required, wherein we divide by their product without checking it's non-zero. Greg Farnum
02:34 PM Bug #14607 (Duplicate): Some fuse tests fail with: failed to remove ‘/home/ubuntu/cephtest’: Dire...
Dupe of: http://tracker.ceph.com/issues/13422 John Spray
07:13 AM Bug #14607: Some fuse tests fail with: failed to remove ‘/home/ubuntu/cephtest’: Directory not empty
I'm seeing this a *bunch* now as well. eg http://pulpito.ceph.com/gregf-2016-02-03_09:19:17-fs-greg-fs-testing-23---b... Greg Farnum
07:56 AM Feature #14640 (New): nfs: qa: evaluate connectathon NFS tests for applicability to our suite
https://fedorapeople.org/cgit/steved/public_git/cthon04.git/tree/README
Is there anything in there that might be u...
Greg Farnum

02/03/2016

09:37 AM Backport #14624 (In Progress): hammer: fsx failed to compile
Nathan Cutler
09:32 AM Backport #14624 (Resolved): hammer: fsx failed to compile
https://github.com/ceph/ceph/pull/7501 Nathan Cutler
08:49 AM Bug #12710 (Resolved): fsstress.sh fails
Nathan Cutler
06:40 AM Bug #10436 (Fix Under Review): ceph-fuse: snapshot flushing from page cache to Client is not cohe...
https://github.com/ceph/ceph/pull/7495 Zheng Yan
05:46 AM Backport #14584 (Resolved): hammer: fsstress.sh fails
Loïc Dachary

02/02/2016

06:32 PM Bug #13903 (Resolved): Failure in TestStrays.test_ops_throttle
Whoops, merged this last week. Greg Farnum
02:29 PM Bug #14608 (Can't reproduce): snaptests.yaml failure: [WRN] open_snap_parents has:" in cluster log

http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-30_18:04:02-fs-master---basic-openstack/13302/
John Spray
02:18 PM Bug #14607 (Duplicate): Some fuse tests fail with: failed to remove ‘/home/ubuntu/cephtest’: Dire...

e.g. http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-30_18:04:02-fs-master---basic-openstack/13319/
& ...
John Spray
01:57 PM Bug #10436 (In Progress): ceph-fuse: snapshot flushing from page cache to Client is not coherent
Zheng Yan
11:44 AM Bug #10436: ceph-fuse: snapshot flushing from page cache to Client is not coherent
Zheng Yan
09:21 AM Bug #14557 (Duplicate): snaps: failed snaptest-multiple-capsnaps.sh
... Zheng Yan

02/01/2016

03:33 AM Backport #14584 (In Progress): hammer: fsstress.sh fails
Loïc Dachary
03:30 AM Backport #14584 (Resolved): hammer: fsstress.sh fails
https://github.com/ceph/ceph/pull/7454 Loïc Dachary
03:26 AM Bug #12710 (Pending Backport): fsstress.sh fails
Loïc Dachary

01/29/2016

10:25 AM Backport #14067 (In Progress): infernalis : Ceph file system is not freeing space
Abhishek Varshney
10:18 AM Backport #14490 (In Progress): infernalis: fsx failed to compile
Abhishek Varshney
05:19 AM Bug #14557 (Duplicate): snaps: failed snaptest-multiple-capsnaps.sh
This is on a testing branch, but it's about to get merged to master.
http://pulpito.ceph.com/gregf-2016-01-26_15:35:...
Greg Farnum
05:15 AM Backport #12350 (In Progress): Provided logrotate setup does not handle ceph-fuse correctly
h3. original description
(move here because Backport issues only have a link to the PR in the description)
OS: ...
Loïc Dachary

01/28/2016

02:50 AM Bug #13926 (Need More Info): lockup in multithreaded application
did not find anything in the log Zheng Yan

01/27/2016

06:01 AM Bug #13926: lockup in multithreaded application
Zheng, anything come out of this? Greg Farnum
05:59 AM Bug #10336 (Can't reproduce): hung ffsb test
Greg Farnum
05:57 AM Bug #13546 (Resolved): mv of directories hung Ceph filesystem
Greg Farnum

01/25/2016

05:10 AM Backport #14490 (Resolved): infernalis: fsx failed to compile
https://github.com/ceph/ceph/pull/7429 Loïc Dachary
04:10 AM Bug #14489 (New): ovh: ENOSPC on multiple_rsync.sh as part of cfuse_workunit_misc
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-16_18:04:01-fs-master---basic-openstack/4272/ Greg Farnum
04:07 AM Bug #14488: ovh: ENOSPC in kernel_untar_build
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-18_20:08:01-kcephfs-master-testing-basic-openstack/6824/ Greg Farnum
03:58 AM Bug #14488 (New): ovh: ENOSPC in kernel_untar_build
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-16_20:08:02-kcephfs-master-testing-basic-openstack/4363/ Greg Farnum
03:50 AM Bug #13977: ovh: sambatorture.scan-pipe_number fails on ENOMEM
http://teuthology.ovh.sepia.ceph.com/teuthology/teuthology-2016-01-16_23:14:01-samba-master---basic-openstack/4432/ Greg Farnum
03:48 AM Bug #14486 (New): ovh: samba test filled up disk
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-17_23:14:01-samba-hammer---basic-openstack/5996/ Greg Farnum

01/21/2016

08:15 PM Feature #14456 (Resolved): mon: prevent older/incompatible clients from mounting the file system
If the kernel client mounts something with quotas set, we should provide server-side warning messages. For now, this ... Greg Farnum
01:32 PM Support #14450 (New): ceph-fuse does not work without loaded ceph/rbd kernel module writing big data
Hi,
i was installing different OS ( centos 6, centos 7, Debian 8 ).
I was installing ceph-fuse only ( hammer r...
Oliver Dzombc
01:32 PM Bug #14365 (Fix Under Review): unsafe handle_config_change() methods
John Spray
01:01 AM Bug #14395: cephfs_journal_tool fails
Sorry~~ I made a mistake xie xingguo

01/20/2016

02:42 PM Bug #13903 (Fix Under Review): Failure in TestStrays.test_ops_throttle
Zheng Yan
02:41 PM Bug #13903: Failure in TestStrays.test_ops_throttle
https://github.com/ceph/ceph/pull/7297 Zheng Yan
02:55 AM Bug #13903: Failure in TestStrays.test_ops_throttle
Zheng, please take a look. Greg Farnum
02:12 PM Bug #14365: unsafe handle_config_change() methods
Also, MDSDaemon::handle_conf_change was broken because it was taking mds_lock, when mds_lock is already taken in the ... John Spray
01:00 PM Bug #14365: unsafe handle_config_change() methods
Turns out there's a pre-existing lock cycle issue similar to #14374 here, we just never noticed because live config c... John Spray
12:58 PM Bug #14365 (In Progress): unsafe handle_config_change() methods
John Spray
11:43 AM Bug #14374 (Fix Under Review): MDS asok handlers trigger lock cycle assertion if they take mds_lock
https://github.com/ceph/ceph/pull/7295 John Spray
02:49 AM Bug #14395 (Resolved): cephfs_journal_tool fails
Greg Farnum
02:24 AM Bug #14377 (Resolved): [ FAILED ] LibCephFS.DirLs
Greg Farnum
02:17 AM Bug #11517 (Resolved): Libcephfs: Doesn't check file's open mode when do read/write
Greg Farnum
02:13 AM Bug #14254 (Resolved): failed pjd chown test 117
Greg Farnum

01/19/2016

07:42 PM Feature #14427 (Resolved): qa: run snapshot tests under thrashing
We run snapshot tests in their own subsuite right now to verify they keep functioning, but we do not test them under ... Greg Farnum
07:41 PM Bug #10436: ceph-fuse: snapshot flushing from page cache to Client is not coherent
Once this is fixed we need to re-enable snaptest-snap-rm-cmp.sh in the snaptests.yaml qa-suite config fragment. Greg Farnum
07:39 PM Feature #3819 (Resolved): mds: re-add snaptests to qa suite
https://github.com/ceph/ceph-qa-suite/pull/678 Greg Farnum
07:34 PM Bug #13903: Failure in TestStrays.test_ops_throttle
http://qa-proxy.ceph.com/teuthology/gregf-2016-01-18_19:56:11-fs-greg-fs-speculative-118---basic-mira/32912/ Greg Farnum
03:01 PM Bug #14255: qa: we are filling smithi disks with ffsb workloads
ffsb (config is at qa/workunits/suites/random_write.32.ffsb) only use about 13G space, no matter how long it runs. I ... Zheng Yan
02:02 AM Bug #14255: qa: we are filling smithi disks with ffsb workloads
I haven't looked yet but I suspect we're just running for a set period of time and the smithis are so much faster tha... Greg Farnum
08:45 AM Bug #14357 (Resolved): Delay in clientreplay on quiet clusters
Zheng Yan

01/18/2016

08:39 PM Bug #14365: unsafe handle_config_change() methods
I think these should be okay, but it's easy to fix. John, can you establish that we don't need locks or else set them... Greg Farnum
08:26 AM Bug #14395 (Fix Under Review): cephfs_journal_tool fails
https://github.com/ceph/ceph-qa-suite/pull/801 Zheng Yan
08:21 AM Bug #14395 (Resolved): cephfs_journal_tool fails
http://qa-proxy.ceph.com/teuthology/teuthology-2016-01-13_14:03:02-fs-jewel---basic-smithi/28198/teuthology.log
<p...
Zheng Yan
07:06 AM Bug #14380 (Fix Under Review): "ceph mds setmap" crashes mon on invalid input
https://github.com/ceph/ceph/pull/7262 Zheng Yan

01/15/2016

02:38 PM Bug #14384 (Pending Backport): fsx failed to compile
I sent a patch upstream and pushed a patch to the workunit in master branch so that it checks out a working commit in... Greg Farnum
01:39 PM Bug #14384 (Resolved): fsx failed to compile
http://teuthology.ovh.sepia.ceph.com/teuthology/teuthology-2016-01-11_23:04:01-fs-master---basic-openstack/2232/
htt...
Zheng Yan
01:12 PM Bug #14379 (Fix Under Review): Add confirmation flag to "ceph mds rmfailed"
https://github.com/ceph/ceph/pull/7248 Zheng Yan
12:30 AM Bug #14379 (Resolved): Add confirmation flag to "ceph mds rmfailed"

It's horribly dangerous but has a rather non-threatening name.
John Spray
09:27 AM Bug #13546 (Fix Under Review): mv of directories hung Ceph filesystem
I added it to https://github.com/ceph/ceph/pull/7199 Zheng Yan
09:22 AM Bug #14377 (Fix Under Review): [ FAILED ] LibCephFS.DirLs
https://github.com/ceph/ceph/pull/7246 Zheng Yan
06:09 AM Bug #14377: [ FAILED ] LibCephFS.DirLs
It's caused by duplicated entries in readdir result. This can happen when readdir requires several mds requests and t... Zheng Yan
12:35 AM Bug #14380 (Resolved): "ceph mds setmap" crashes mon on invalid input
Needs exception handling around MDSMap::decode John Spray

01/14/2016

07:27 PM Feature #13569 (Resolved): ceph-fuse: support direct IO
Greg Farnum
07:25 PM Bug #13546 (In Progress): mv of directories hung Ceph filesystem
Do you have a PR or any other commits to go with that? Is it safe to unilaterally not drop caps in a case like this? Greg Farnum
06:18 PM Bug #14377: [ FAILED ] LibCephFS.DirLs
Zheng, you've been fixing a lot here lately, please take a look! Greg Farnum
06:18 PM Bug #14377 (Resolved): [ FAILED ] LibCephFS.DirLs
http://pulpito.ceph.com/gregf-2016-01-12_23:23:33-fs-greg-fs-testing-1-12-1---basic-mira/26500/
This failed the Di...
Greg Farnum
03:41 PM Bug #14374 (Resolved): MDS asok handlers trigger lock cycle assertion if they take mds_lock

http://pulpito.ceph.com/gregf-2016-01-12_23:29:42-fs-greg-fs-speculative---basic-mira/26556
The asok handler is ...
John Spray
03:14 AM Bug #14319: Double decreased the count to trim caps which will cause failing to respond to cache ...
This issue also exists on master, so close original PR and create this new one.
https://github.com/ceph/ceph/pull/...
Zhi Zhang
02:39 AM Backport #12350 (Pending Backport): Provided logrotate setup does not handle ceph-fuse correctly
Zhi Zhang
12:42 AM Bug #14365 (Resolved): unsafe handle_config_change() methods
The handle_conf_change() methods in these files look potentially unsafe (modifying data without locks):... Josh Durgin

01/13/2016

02:21 PM Bug #13903: Failure in TestStrays.test_ops_throttle
If anyone has time, yes -- given enough time I can figure it out but it might be more obvious to someone more familia... John Spray
01:11 AM Bug #13903: Failure in TestStrays.test_ops_throttle
I think you talked about this in standup but I'm forgetting — do you need somebody else to look over the caps stuff h... Greg Farnum
02:20 PM Bug #13546: mv of directories hung Ceph filesystem
I have an explanation for this.

see https://github.com/ukernel/ceph/commit/a750c361cd631a1f87ee152083d2a42c49fd02b6
Zheng Yan
02:19 AM Bug #13546: mv of directories hung Ceph filesystem
Bumping this up as we ought to examine it and I think it's been lost in the shuffle. Greg Farnum
12:05 PM Bug #14357 (Fix Under Review): Delay in clientreplay on quiet clusters
https://github.com/ceph/ceph/pull/7216 John Spray
10:38 AM Bug #14357 (Resolved): Delay in clientreplay on quiet clusters
Because we are checking for clientreplay_done at the end of _dispatch, if a request is completing via a commit contex... John Spray
08:01 AM Bug #11517 (Fix Under Review): Libcephfs: Doesn't check file's open mode when do read/write
https://github.com/ceph/ceph/pull/7209 Zheng Yan
02:24 AM Bug #14256 (Resolved): mds: objecter assert on shutdown
Sage Weil

01/12/2016

04:03 PM Bug #14195: test_full_fclose fails (tasks.cephfs.test_full.TestClusterFull)

So there are two different failure modes here, we either get the exception out of the fsync() or out of the followi...
John Spray
02:37 PM Bug #14195 (In Progress): test_full_fclose fails (tasks.cephfs.test_full.TestClusterFull)
John Spray
10:47 AM Bug #14195: test_full_fclose fails (tasks.cephfs.test_full.TestClusterFull)
The symptom is that we're getting ENOSPC from write() calls during buffered IO, where we should be getting them from ... John Spray
10:29 AM Bug #14195: test_full_fclose fails (tasks.cephfs.test_full.TestClusterFull)
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-11_23:04:01-fs-master---basic-openstack/2263/ John Spray
02:43 PM Bug #14144 (Resolved): standy-replay MDS does not cleanup finished replay threads
Greg Farnum
02:39 PM Bug #11783 (Resolved): protocol: flushing caps on MDS restart can go bad
Greg Farnum
02:39 PM Bug #11517: Libcephfs: Doesn't check file's open mode when do read/write
Oh, this might not be done for read-only fds. Need to check. Greg Farnum
02:37 PM Backport #12350 (Fix Under Review): Provided logrotate setup does not handle ceph-fuse correctly
Greg Farnum
02:37 PM Bug #14258 (In Progress): qa: failed test_full_fsync
John Spray
09:19 AM Bug #14254 (Fix Under Review): failed pjd chown test 117
https://github.com/ceph/ceph/pull/7199 Zheng Yan

01/11/2016

11:17 PM Feature #14146 (Resolved): MDS: expose state of boot/replay to admins
John Spray
08:11 PM Bug #13903: Failure in TestStrays.test_ops_throttle
This is reproducible with a simpler "delete lots of files and then their directory" test https://github.com/ceph/ceph... John Spray
05:13 PM Bug #13903: Failure in TestStrays.test_ops_throttle
The client is receiving a client_caps message for the dir just *after* it's done the unlink. I think that's preventi... John Spray
03:28 PM Bug #13903: Failure in TestStrays.test_ops_throttle
http://pulpito.ceph.com/teuthology-2015-11-23_23:04:04-fs-master---basic-multi/1157802/
In this case I can see the...
John Spray
02:50 PM Bug #13903: Failure in TestStrays.test_ops_throttle

So in all three cases we're seeing just a single inode that's failing to get purged, probably the dir.
http://pu...
John Spray
09:39 AM Bug #14254: failed pjd chown test 117
I interpret this differently.
* client creates inode, mds replies unsafe
* client requests inode change gid 655...
Zheng Yan
05:55 AM Bug #14319: Double decreased the count to trim caps which will cause failing to respond to cache ...
-https://github.com/ceph/ceph/pull/7172- Zhi Zhang
05:48 AM Bug #14319 (Resolved): Double decreased the count to trim caps which will cause failing to respon...
When reaching mds cache size, mds will ask clients to trim its own caps. In ceph-fuse, it will recalculate current ca... Zhi Zhang

01/08/2016

02:26 PM Bug #13903 (In Progress): Failure in TestStrays.test_ops_throttle
John Spray
08:06 AM Bug #14258: qa: failed test_full_fsync
Greg Farnum wrote:
>
> It's obviously *supposed* to be filling things up, but maybe at only 142MB of storage it's ...
Zheng Yan

01/07/2016

09:25 PM Bug #14256 (Fix Under Review): mds: objecter assert on shutdown
I don't know if we ever used "Pending upstream" before, usually when a PR is outstanding we use "Needs review" John Spray
09:24 PM Bug #14256: mds: objecter assert on shutdown
Link to the pull request for convenience:
https://github.com/ceph/ceph/pull/7151
John Spray
07:29 PM Bug #14256: mds: objecter assert on shutdown
Pushed upstream as:... Adam Emerson
06:56 PM Bug #14256: mds: objecter assert on shutdown
This patch should fix it. I'll run make check and push. Adam Emerson
06:12 PM Bug #14256 (In Progress): mds: objecter assert on shutdown
Adam Emerson
03:38 PM Bug #14256: mds: objecter assert on shutdown
Hmm, looks like Objecter is making the assumption that if tick_event is set then it must also be in ceph_timer::event... John Spray
03:05 PM Bug #14257 (Resolved): test_reconnect_timeout failed
Fix merged into ceph-qa-suite master. John Spray
11:26 AM Bug #14257 (Fix Under Review): test_reconnect_timeout failed
https://github.com/ceph/ceph-qa-suite/pull/785 John Spray
11:16 AM Bug #14257 (In Progress): test_reconnect_timeout failed
John Spray
11:16 AM Bug #14257: test_reconnect_timeout failed
Test bug, Filesystem.wait_for_state is counting elapsed time as the number of times it goes through its polling loop ... John Spray
02:05 AM Bug #14254: failed pjd chown test 117
Okay, so the only way we can get into this trouble is if:
1) the inode isn't found prior to replay (ie, it got creat...
Greg Farnum
01:36 AM Bug #14254: failed pjd chown test 117
https://github.com/ceph/ceph/pull/7136 addresses the need to flush Greg Farnum
01:29 AM Bug #14254: failed pjd chown test 117
Okay, this is running into code from the uid/gid enforcement stuff. From https://github.com/ceph/ceph/commit/1957aedd... Greg Farnum
01:14 AM Bug #14254: failed pjd chown test 117
Okay, I believe the important sequence is:
* client creates inode, mds replies unsafe
* client requests inode cha...
Greg Farnum

01/06/2016

09:56 PM Bug #13583 (Resolved): Client::_fsync() on a given file does not wait unsafe requests that create...
Merged to master rather than jewel due to size of patch series and rarity of issues. Greg Farnum
06:39 AM Bug #13583: Client::_fsync() on a given file does not wait unsafe requests that create/modify the...
Causing failures like http://pulpito.ovh.sepia.ceph.com:8081/gregf-2015-12-23_05:34:31-fs-master---basic-openstack/50... Greg Farnum
09:52 PM Bug #14196 (Resolved): test_object_deletion fails (tasks.cephfs.test_damage.TestDamage)
a0b365b220134badcd212bf22386b6124e4dfa69 Greg Farnum
06:26 AM Bug #14196: test_object_deletion fails (tasks.cephfs.test_damage.TestDamage)
Also seen at http://pulpito.ceph.com/gregf-2015-12-21_23:08:59-fs-master---basic-smithi/1795/ Greg Farnum
09:47 PM Feature #14271: directory listing: do not reset when fragmenting
Once completed, update the DirLs test to check order again. Probably just by reverting e20ef4b27869d8eaf1989c2f057c2d... Greg Farnum
09:44 PM Feature #14271 (Resolved): directory listing: do not reset when fragmenting
Right now, if a directory gets fragmented while we're listing (ie, have a dir pointer), we reset the pointer and star... Greg Farnum
09:45 PM Bug #13364 (Resolved): LibCephFS locking tests are failing and/or lockdep asserting
Merged to jewel branch as of 088b8c2c881c8561b9bc96934e04d030e8de9255.
I created a new feature ticket for the direct...
Greg Farnum
06:36 AM Bug #13903: Failure in TestStrays.test_ops_throttle
Here are some correct logs on master:
http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2015-12-21_23:04:02-fs-mas...
Greg Farnum
06:34 AM Bug #14258 (Duplicate): qa: failed test_full_fsync
http://pulpito.ceph.com/gregf-2015-12-21_23:08:59-fs-master---basic-smithi/1822/
http://pulpito.ceph.com/gregf-2015-...
Greg Farnum
06:30 AM Bug #14257: test_reconnect_timeout failed
Haven't looked into this at all but I wonder if it's failing to account for all the clients pinging back early, or th... Greg Farnum
06:28 AM Bug #14257 (Resolved): test_reconnect_timeout failed
http://pulpito.ceph.com/gregf-2015-12-21_23:08:59-fs-master---basic-smithi/1789/... Greg Farnum
06:18 AM Bug #14256 (Resolved): mds: objecter assert on shutdown
http://pulpito.ceph.com/gregf-2015-12-21_23:08:59-fs-master---basic-smithi/1782/
Only saw this once so far and it ...
Greg Farnum
06:15 AM Bug #14255 (New): qa: we are filling smithi disks with ffsb workloads
http://pulpito.ceph.com/gregf-2016-01-04_11:49:51-fs-master---basic-smithi/13310/
http://pulpito.ceph.com/gregf-2016...
Greg Farnum
05:59 AM Bug #14254 (Resolved): failed pjd chown test 117
http://pulpito.ceph.com/gregf-2016-01-04_11:47:54-fs-master---basic-mira/13222/... Greg Farnum
 

Also available in: Atom