Project

General

Profile

Activity

From 07/18/2016 to 08/16/2016

08/16/2016

01:12 PM Bug #16608 (Pending Backport): bashism in src/rbdmap
Jason Dillaman
01:10 PM Bug #16529 (Pending Backport): "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-d...
https://github.com/ceph/ceph/pull/10714 is merged. It should be backported to jewel before https://github.com/ceph/ce... Mykola Golub
01:06 PM Bug #17030 (Pending Backport): The "request lock" RPC message might be incorrectly ignored
Jason Dillaman
12:58 PM Bug #16967 (Pending Backport): rbd bench-write: seg fault when "--io-size" is larger than image s...
Mykola Golub
12:57 PM Bug #16984 (Pending Backport): Disabling pool mirror mode with registered peers results orphaned ...
Mykola Golub
12:55 PM Bug #16740 (Pending Backport): Cannot disable journaling or remove non-mirrored, "non-primary" image
Mykola Golub

08/15/2016

09:58 PM Bug #17030 (Fix Under Review): The "request lock" RPC message might be incorrectly ignored
*PR*: https://github.com/ceph/ceph/pull/10735 Jason Dillaman
08:50 PM Bug #17030 (Resolved): The "request lock" RPC message might be incorrectly ignored
If the lock is being released, the request lock RPC message will be incorrectly dropped on the floor w/o an acknowled... Jason Dillaman
09:56 PM Bug #16600 (Resolved): librbd: fix missing return statement if failed to get mirror image state
Loïc Dachary
09:28 PM Backport #16601 (Resolved): jewel: librbd: fix missing return statement if failed to get mirror i...
Loïc Dachary
07:54 PM Cleanup #17028 (New): rbd-mirror: new test case for multiple one-way peers
The rbd-mirror daemon does not currently support a mesh topology of peers, but it should currently be able to support... Jason Dillaman
07:48 PM Subtask #17015 (In Progress): rbd-mirror HA: cls_lock needs the ability to update lock cookie
Jason Dillaman
06:23 PM Bug #16608 (Fix Under Review): bashism in src/rbdmap
*PR*: https://github.com/ceph/ceph/pull/10733 Jason Dillaman
06:07 PM Bug #16608 (In Progress): bashism in src/rbdmap
Jason Dillaman
02:01 PM Backport #16701: jewel: rbd-mirror: image sync throttle needs to use pool id + image id to form u...
... Loïc Dachary
01:44 PM Backport #17026 (Resolved): jewel: doc: format 2 now is the default image format
https://github.com/ceph/ceph/pull/10732 Loïc Dachary

08/12/2016

07:36 PM Feature #12901 (Resolved): librbd: zone + zonegroup configuration
Jason Dillaman
07:33 PM Feature #14410: snapshots need to support internal vs external use cases
This would also be helpful for consistency group snapshots. Jason Dillaman
07:32 PM Subtask #14417 (Resolved): Interlock RBD mirroring daemon and acquiring exclusive lock
Jason Dillaman
07:30 PM Subtask #14739 (Can't reproduce): Support deleting a journal when its being replayed by the RBD m...
Jason Dillaman
07:30 PM Feature #11766 (Resolved): rbd-mirror: basic structure
Jason Dillaman
07:25 PM Feature #4389 (Closed): get md5 of image
You can run 'rbd export <image-spec> - | <favorite checksum command>". Jason Dillaman
07:19 PM Feature #2231 (Resolved): librbd: expose header change (resize?) via api
Included in commit 9952b75a0c8c9bed9792cb77671fd5588f32630f Jason Dillaman
07:11 PM Subtask #17020 (Resolved): rbd-mirror HA: pool replayer should be started/stopped when lock acqui...
This will ensure that only a single rbd-mirror process is replicating a single pool. Jason Dillaman
07:09 PM Subtask #17019 (Resolved): rbd-mirror HA: create pool locker / leader class
This would establish the watch on the rbd_mirror_leader object, attempt to lock it, watch for heartbeats if the lock ... Jason Dillaman
07:07 PM Subtask #17018 (Resolved): rbd-mirror HA: add new lock released/acquired and heartbeat messages
Need to decide if these notifications should be sent against existing rbd_mirroring object or new rbd_mirror_leader o... Jason Dillaman
07:01 PM Subtask #17017 (Resolved): rbd-mirror HA: move librbd::image_watcher::Notifier to librbd::object_...
Need a generic asynchronous watch/notification framework Jason Dillaman
06:59 PM Subtask #17016 (Resolved): rbd-mirror HA: create a generic/re-usable managed lock class
This should be based on the existing librbd ExclusiveLock class (and associated namespace helpers). It should support... Jason Dillaman
06:56 PM Subtask #17015 (Resolved): rbd-mirror HA: cls_lock needs the ability to update lock cookie
In order to support blacklisting a client, the watch cookie is stored encoded within the lock cookie. Need a way to u... Jason Dillaman
06:54 PM Feature #17014 (Resolved): rbd-mirror: add support for active/passive daemon instances
See http://pad.ceph.com/p/rbd_mirror_scale Jason Dillaman
06:50 PM Bug #16773 (In Progress): FAILED assert(m_image_ctx.journal == nullptr)
Jason Dillaman
06:50 PM Bug #16529 (Fix Under Review): "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-d...
*PR*: https://github.com/ceph/ceph/pull/10714
*QA suite PR*: https://github.com/ceph/ceph-qa-suite/pull/1116
Jason Dillaman
06:40 PM Bug #16529 (In Progress): "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-distro...
Jason Dillaman
06:19 PM Bug #16898 (Fix Under Review): "TestLibRBD.UpdateFeatures" tests failed in upgrade:client-upgrade...
*PR*: https://github.com/ceph/ceph/pull/10713 Jason Dillaman
06:18 PM Bug #16898 (In Progress): "TestLibRBD.UpdateFeatures" tests failed in upgrade:client-upgrade-infe...
Jason Dillaman
06:18 PM Feature #17012: generate a UUID for image HEAD and each snapshot at image/snapshot creation time
Should probably add tune2fs -U functionality as well. Ilya Dryomov
05:22 PM Feature #17012 (New): generate a UUID for image HEAD and each snapshot at image/snapshot creation...
See http://marc.info/?l=ceph-devel&m=147101148114069&w=2 for context.
The only question here is old images. May n...
Ilya Dryomov
04:41 PM Bug #16740 (Fix Under Review): Cannot disable journaling or remove non-mirrored, "non-primary" image
*PR*: https://github.com/ceph/ceph/pull/10712 Jason Dillaman
03:45 PM Feature #17010 (Resolved): RBD default features should be negotiated with the OSD
For example, attempting to create an image with a Jewel librbd client against a Hammer OSD will result in failure sin... Jason Dillaman
01:58 PM Bug #16984 (Fix Under Review): Disabling pool mirror mode with registered peers results orphaned ...
*PR*: https://github.com/ceph/ceph/pull/10711 Jason Dillaman
12:30 PM Bug #16984 (In Progress): Disabling pool mirror mode with registered peers results orphaned mirro...
Jason Dillaman
12:28 PM Bug #16967 (Fix Under Review): rbd bench-write: seg fault when "--io-size" is larger than image s...
*PR*: https://github.com/ceph/ceph/pull/10708 Jason Dillaman
12:19 PM Bug #16967 (In Progress): rbd bench-write: seg fault when "--io-size" is larger than image size
Jason Dillaman
09:53 AM Backport #17008 (Closed): jewel: Support asynchronous v2 image creation/cloning
Loïc Dachary
09:48 AM Backport #17005 (Resolved): jewel: ImageReplayer::is_replaying does not include flush state
https://github.com/ceph/ceph/pull/10790 Loïc Dachary
09:48 AM Backport #17004 (Resolved): jewel: rbd-mirror: FAILED assert(m_state == STATE_STOPPING)
https://github.com/ceph/ceph/pull/10791 Loïc Dachary
01:38 AM Bug #16855 (Fix Under Review): rbd mirror: after promote, the mirror image often be up+error
*PR*: https://github.com/ceph/ceph/pull/10703 Jason Dillaman
12:18 AM Bug #17000 (Duplicate): "[ FAILED ] TestLibRBD.LockingPP" in upgrade:hammer-x-infernalis-distro...
Run: http://pulpito.ceph.com/teuthology-2016-08-11_18:05:01-upgrade:hammer-x-infernalis-distro-basic-vps/
Job: 35860...
Yuri Weinstein

08/11/2016

07:09 PM Documentation #16999 (Resolved): RBD quick start guide will fail due to default image features
http://docs.ceph.com/docs/master/start/quick-rbd/
The documentation should be updated to only enable the layering ...
Jason Dillaman
03:36 PM Feature #15321: Support asynchronous v2 image creation/cloning
Holding this for backport to jewel until required by #15764 Jason Dillaman
03:20 PM Feature #15321: Support asynchronous v2 image creation/cloning
Currently there's support only for async create. Venky Shankar
03:19 PM Feature #15321 (Pending Backport): Support asynchronous v2 image creation/cloning
Venky Shankar
02:56 PM Backport #16904 (In Progress): jewel: journal should prefetch small chunks of the object during r...
Jason Dillaman
02:31 PM Backport #16511 (In Progress): jewel: rbd-mirror: 'wait_for_scheduled_deletion' callback might de...
Jason Dillaman
02:27 PM Backport #16512 (In Progress): jewel: rbd-mirror: potential race condition when restarting image ...
Jason Dillaman
02:22 PM Backport #16747 (In Progress): jewel: rbd-mirror: snap rename does not correctly replicate
Jason Dillaman
02:10 PM Backport #16735 (In Progress): jewel: rbd-nbd does not properly handle resize notifications
Jason Dillaman
02:05 PM Backport #16701 (In Progress): jewel: rbd-mirror: image sync throttle needs to use pool id + imag...
Jason Dillaman
01:53 PM Backport #16658 (In Progress): jewel: rbd-mirror: gracefully handle being blacklisted
Jason Dillaman
01:04 PM Bug #16980 (Pending Backport): rbd-mirror: FAILED assert(m_state == STATE_STOPPING)
Mykola Golub
06:16 AM Bug #16970 (Pending Backport): ImageReplayer::is_replaying does not include flush state
Mykola Golub
03:09 AM Bug #16773: FAILED assert(m_image_ctx.journal == nullptr)
If the rbd image header watch is lost ...... Jason Dillaman

08/10/2016

08:55 PM Bug #16773: FAILED assert(m_image_ctx.journal == nullptr)
Repeated here: http://pulpito.ceph.com/jdillaman-2016-08-08_08:12:47-rbd-wip-jd-testing---basic-vps/354673/
Both c...
Jason Dillaman
07:32 PM Cleanup #16991 (Resolved): rbd-mirror split-brain issues should be clearly visible in mirror status
Ensure that when a split-brain is detected, the image is moved to the error state under "rbd mirror image status" and... Jason Dillaman
07:29 PM Cleanup #16990 (Resolved): 'rbd image-meta remove' of missing key does not return error
The command should indicate that the operation failed if the image or key doesn't exist. Jason Dillaman
07:12 PM Cleanup #16989 (New): 'rbd mirror pool' commands should report error if action not executed
Jason Dillaman
07:11 PM Cleanup #16989 (Duplicate): 'rbd mirror pool' commands should report error if action not executed
Jason Dillaman
07:11 PM Cleanup #16989 (New): 'rbd mirror pool' commands should report error if action not executed
Some examples:
1. rbd mirror pool enable <PoolName> pool :- Doesn't throw a warning message when the pool is alre...
Jason Dillaman
07:06 PM Bug #16988 (Resolved): 'rbd feature disable' against non-primary image results in hung CLI
The command will continuously repeat an error about requesting the exclusive lock:... Jason Dillaman
07:00 PM Cleanup #16987 (Resolved): 'rbd du' of missing image does not return error
The command should indicate that the operation failed if the image doesn't exist. Jason Dillaman
06:58 PM Cleanup #16986 (New): 'rbd mirror pool peer remove' of missing peer does not return error
The command should indicate that the operation failed if the peer doesn't exist. Jason Dillaman
06:52 PM Cleanup #16985 (Resolved): Improve error reporting from "rbd feature enable/disable"
For example, attempting to enable the journaling feature on an image that doesn't have the exclusive-lock feature ena... Jason Dillaman
06:36 PM Bug #16984 (Resolved): Disabling pool mirror mode with registered peers results orphaned mirrored...
Attempting to disable mirroring while one or more peers are still registered correctly results in an -EBUSY error. Ho... Jason Dillaman
06:20 PM Bug #16980 (Fix Under Review): rbd-mirror: FAILED assert(m_state == STATE_STOPPING)
*PR*: https://github.com/ceph/ceph/pull/10667 Jason Dillaman
01:13 PM Bug #16980 (Resolved): rbd-mirror: FAILED assert(m_state == STATE_STOPPING)
If the journal contains a failed maintenance op followed immediately by a new exclusive lock epoch (e.g. a second mai... Jason Dillaman
11:37 AM Bug #16741 (Resolved): io getting stuck after advancing journal object set
Yes, it is not reproduced any more, after #16223 was resolved. Mykola Golub
11:29 AM Bug #16478: Non-primary image is recording journal events during image sync
PR: https://github.com/ceph/ceph/pull/10462 Mykola Golub
08:37 AM Bug #16478: Non-primary image is recording journal events during image sync
What is the corresponding pull request ? Loïc Dachary
08:54 AM Backport #16658: jewel: rbd-mirror: gracefully handle being blacklisted
https://github.com/ceph/ceph/commit/48f301decbf1f27937bb77a3b47e54933f272d7d conflicts Loïc Dachary
08:52 AM Backport #16978: jewel: rbd-mirror: FAILED assert(m_on_update_status_finish == nullptr)
https://github.com/ceph/ceph/commit/6a465d9dad417e8b52909c5478f7e3e433748948 conflicts Loïc Dachary
07:03 AM Backport #16978 (Resolved): jewel: rbd-mirror: FAILED assert(m_on_update_status_finish == nullptr)
https://github.com/ceph/ceph/pull/10792 Loïc Dachary
08:49 AM Backport #16951 (In Progress): jewel: ceph 10.2.2 rbd status on image format 2 returns "(2) No su...
Loïc Dachary
08:43 AM Backport #16950: jewel: librbd/ExclusiveLock.cc: 197: FAILED assert(m_watch_handle != 0)
https://github.com/ceph/ceph/pull/10574/commits/814c305ce8c35b5ce01d7e29a912d5ef3978754b conflicts Loïc Dachary
08:39 AM Backport #16904: jewel: journal should prefetch small chunks of the object during replay
https://github.com/ceph/ceph/commit/dad8328f2d502d18923c35f7b86a0cc2ccec133a conflicts Loïc Dachary
08:37 AM Backport #16903 (Need More Info): jewel: Non-primary image is recording journal events during ima...
Loïc Dachary
08:35 AM Backport #16902: jewel: rbd-mirror: image deleter should use pool id + global image uuid for key
https://github.com/ceph/ceph/commit/25203a8a9d59ff025d223ec1afaeb14946d54993 conflicts Loïc Dachary
08:30 AM Backport #16869 (In Progress): jewel: Discard hangs when 'rbd_skip_partial_discard' is enabled
Loïc Dachary
08:25 AM Backport #16868 (In Progress): jewel: Prevent the creation of a clone from a non-primary mirrored...
Loïc Dachary
08:24 AM Backport #16867 (In Progress): jewel: mkfs.xfs slow performance with discards and object map
Loïc Dachary
08:23 AM Backport #16796 (In Progress): jewel: Renaming old format image results in "Transport endpoint is...
Loïc Dachary
08:22 AM Backport #16747: jewel: rbd-mirror: snap rename does not correctly replicate
https://github.com/ceph/ceph/commit/2f4cb26d8bbd1457bc261547103b56ad40b3c464 cherry-picks ok, but https://github.com/... Loïc Dachary
08:19 AM Backport #16735: jewel: rbd-nbd does not properly handle resize notifications
https://github.com/ceph/ceph/commit/9952b75a0c8c9bed9792cb77671fd5588f32630f conflicts Loïc Dachary
08:18 AM Backport #16701: jewel: rbd-mirror: image sync throttle needs to use pool id + image id to form u...
https://github.com/ceph/ceph/commit/8ad36cab7cbf3492bfa972b43e4a5f75a110bfe6 conflicts Loïc Dachary
08:15 AM Backport #16593 (In Progress): jewel: FAILED assert(object_no < m_object_map.size())
Loïc Dachary
08:13 AM Backport #16576 (In Progress): jewel: rbd-mirror: FAILED assert(m_local_image_ctx->object_map != ...
Loïc Dachary
08:07 AM Backport #16513 (In Progress): jewel: rbd-mirror: image-replayer: segfault when removing resync l...
Loïc Dachary
08:03 AM Backport #16458 (In Progress): jewel: Potential crash during journal::Replay shut down
Loïc Dachary
08:02 AM Backport #16373 (In Progress): jewel: rbd-mirror: gracefully handle missing sync point snapshots
Loïc Dachary
07:21 AM Bug #16471 (Resolved): Image removal doesn't necessarily clean up all rbd_mirroring entries
Loïc Dachary
07:21 AM Bug #16470 (Resolved): Timeout sending mirroring notification shouldn't result in failure
Loïc Dachary
07:21 AM Bug #16450 (Resolved): Close journal and object map before flagging exclusive lock as released
Loïc Dachary
07:21 AM Bug #16446 (Resolved): ExclusiveLock object leaked when switching to snapshot
Loïc Dachary
07:21 AM Bug #16445 (Resolved): Whitelist EBUSY error from "snap unprotect" for journal replay
Loïc Dachary
07:21 AM Bug #16411 (Resolved): rbd-mirror should disable proxied maintenance ops for non-primary image
Loïc Dachary
07:21 AM Bug #16364 (Resolved): Crash when utilizing advisory locking API functions
Loïc Dachary
07:21 AM Bug #16363 (Resolved): Journal duplicate op detection can cause lockdep error
Loïc Dachary
07:21 AM Bug #16362 (Resolved): Journal needs to handle duplicate maintenance op tids
Loïc Dachary
07:21 AM Bug #16352 (Resolved): rbd-mirror: ensure replay status formatter has completed before stopping r...
Loïc Dachary
07:21 AM Bug #16350 (Resolved): Object map/fast-diff invalidated if journal replays the same snap remove e...
Loïc Dachary
07:21 AM Bug #16348 (Resolved): Unable to disable journaling feature if in unexpected mirror state
Loïc Dachary
07:21 AM Bug #16230 (Resolved): rbd-mirror: potential race condition accessing local image journal
Loïc Dachary
07:21 AM Bug #16198 (Resolved): Possible race condition during journal transition from replay to ready
Loïc Dachary
07:21 AM Bug #15761 (Resolved): When journaling is enabled, a flush request shouldn't flush the cache
Loïc Dachary

08/09/2016

07:39 PM Bug #16974 (Resolved): rbd-mirror: force-promoted image will remain R/O until local rbd-mirror da...
If the local rbd-mirror daemon was replicating events from a remote primary image to a local non-primary image and an... Jason Dillaman
01:19 PM Bug #16956 (Pending Backport): rbd-mirror: FAILED assert(m_on_update_status_finish == nullptr)
Mykola Golub
12:14 PM Bug #16970 (Fix Under Review): ImageReplayer::is_replaying does not include flush state
*PR*: https://github.com/ceph/ceph/pull/10627 Jason Dillaman
12:06 PM Bug #16970 (Resolved): ImageReplayer::is_replaying does not include flush state
Jason Dillaman
06:45 AM Backport #16484 (Resolved): jewel: ExclusiveLock object leaked when switching to snapshot
Loïc Dachary
06:31 AM Backport #16315 (Resolved): jewel: When journaling is enabled, a flush request shouldn't flush th...
Loïc Dachary
06:31 AM Backport #16371 (Resolved): jewel: rbd-mirror: ensure replay status formatter has completed befor...
Loïc Dachary
06:31 AM Backport #16372 (Resolved): jewel: Unable to disable journaling feature if in unexpected mirror s...
Loïc Dachary
06:31 AM Backport #16423 (Resolved): jewel: Journal duplicate op detection can cause lockdep error
Loïc Dachary
06:31 AM Backport #16424 (Resolved): jewel: Journal needs to handle duplicate maintenance op tids
Loïc Dachary
06:31 AM Backport #16425 (Resolved): jewel: rbd-mirror: potential race condition accessing local image jou...
Loïc Dachary
06:31 AM Backport #16426 (Resolved): jewel: Possible race condition during journal transition from replay ...
Loïc Dachary
06:31 AM Backport #16459 (Resolved): jewel: rbd-mirror should disable proxied maintenance ops for non-prim...
Loïc Dachary
06:31 AM Backport #16460 (Resolved): jewel: Crash when utilizing advisory locking API functions
Loïc Dachary
06:31 AM Backport #16482 (Resolved): jewel: Timeout sending mirroring notification shouldn't result in fai...
Loïc Dachary
06:31 AM Backport #16483 (Resolved): jewel: Close journal and object map before flagging exclusive lock as...
Loïc Dachary
06:30 AM Backport #16485 (Resolved): jewel: Whitelist EBUSY error from "snap unprotect" for journal replay
Loïc Dachary
06:30 AM Backport #16486 (Resolved): jewel: Object map/fast-diff invalidated if journal replays the same s...
Loïc Dachary
06:30 AM Backport #16514 (Resolved): jewel: Image removal doesn't necessarily clean up all rbd_mirroring e...
Loïc Dachary
02:06 AM Bug #16394 (Rejected): Ceph RBD
after offline talk, it's user's code problem. Haomai Wang
12:32 AM Bug #16394: Ceph RBD
@Junming: ping Jason Dillaman
12:30 AM Bug #16179: rbd-mirror: image sync object map reload logs message
Jason Dillaman
12:30 AM Bug #16741 (Need More Info): io getting stuck after advancing journal object set
Jason Dillaman
12:27 AM Bug #16967 (Resolved): rbd bench-write: seg fault when "--io-size" is larger than image size
Jason Dillaman

08/08/2016

10:05 PM Subtask #15239 (Resolved): Throttle in-flight image syncs to only a X concurrent
*PR*: https://github.com/ceph/ceph/pull/9623 Jason Dillaman
10:02 PM Subtask #14414 (Closed): Add new "exclusive lock released" journal event to librbd
Resolved issue by tracking demotion/promotion via journal tags. Jason Dillaman
09:59 PM Subtask #15108 (Resolved): Periodically update the sync point object number during sync
*PR*: https://github.com/ceph/ceph/pull/9699 Jason Dillaman
08:01 PM Bug #16962 (Resolved): rbd-mirror: snap protect of non-layered image results in split-brain
Attempting to protect a snapshot against an image that doesn't support layering results in an error:... Jason Dillaman
07:57 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
@Kjetil: unfortunately we cannot just increase the length rbd_image_info_t::block_name_prefix [1] since that would br... Jason Dillaman
05:31 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
Reading into this - librbd/rbd having generated slightly out-of-spec id's is our mess to deal with ? (We get to maint... Kjetil Joergensen
06:42 PM Bug #16956 (Fix Under Review): rbd-mirror: FAILED assert(m_on_update_status_finish == nullptr)
*PR*: https://github.com/ceph/ceph/pull/10613 Jason Dillaman
02:38 PM Bug #16956 (Resolved): rbd-mirror: FAILED assert(m_on_update_status_finish == nullptr)
Jason Dillaman
01:25 PM Bug #16855 (In Progress): rbd mirror: after promote, the mirror image often be up+error
Actually, I was able to repeat the issue. Thanks. Jason Dillaman
01:09 PM Bug #16855: rbd mirror: after promote, the mirror image often be up+error
@de Ian: it doesn't look like debugging was enabled for the first log. Can you please provide the exact steps you per... Jason Dillaman
01:05 PM Bug #16179 (In Progress): rbd-mirror: image sync object map reload logs message
Jason Dillaman
01:00 PM Feature #15706 (Resolved): Optionally limit the maximum number of snapshots for an image
Douglas Fuller
12:46 PM Bug #16176 (Need More Info): objectmap does not show object existence correctly
@Xinxin: I've tried to repeat your findings without success. Can you repeat with "debug rbd = 20" in your ceph client... Jason Dillaman
08:42 AM Bug #16519 (Resolved): librbd: potential use after free on refresh error
Loïc Dachary
08:42 AM Bug #16517 (Resolved): TaskFinisher: cancel all tasks wait until finisher done
Loïc Dachary
08:33 AM Bug #15225 (Resolved): Linking to -lrbd causes process startup times to balloon
Loïc Dachary
08:33 AM Bug #15121 (Resolved): Protect against excessively large object map sizes
Loïc Dachary
08:31 AM Backport #16952 (Resolved): hammer: ceph 10.2.2 rbd status on image format 2 returns "(2) No such...
https://github.com/ceph/ceph/pull/10987 Loïc Dachary
08:31 AM Backport #16951 (Resolved): jewel: ceph 10.2.2 rbd status on image format 2 returns "(2) No such ...
https://github.com/ceph/ceph/pull/10652 Loïc Dachary
08:31 AM Backport #16950 (Resolved): jewel: librbd/ExclusiveLock.cc: 197: FAILED assert(m_watch_handle != 0)
https://github.com/ceph/ceph/pull/10827 Loïc Dachary
08:28 AM Backport #15359 (Rejected): infernalis: Linking to -lrbd causes process startup times to balloon
Loïc Dachary
08:28 AM Backport #15128 (Rejected): infernalis: Protect against excessively large object map sizes
Loïc Dachary
08:20 AM Backport #16518 (Resolved): jewel: TaskFinisher: cancel all tasks wait until finisher done
Loïc Dachary
08:20 AM Backport #16520 (Resolved): jewel: librbd: potential use after free on refresh error
Loïc Dachary

08/07/2016

10:35 AM Bug #16923 (Pending Backport): librbd/ExclusiveLock.cc: 197: FAILED assert(m_watch_handle != 0)
Mykola Golub
10:34 AM Bug #16887 (Pending Backport): ceph 10.2.2 rbd status on image format 2 returns "(2) No such file...
Mykola Golub

08/06/2016

02:09 PM Bug #16176 (In Progress): objectmap does not show object existence correctly
Jason Dillaman
02:08 PM Bug #16019 (Resolved): Failure in TestJournalReplay.Rename after injected socket failure
Believe this is resolved by issue #16404 Jason Dillaman
02:07 PM Bug #15947 (Resolved): Sporadic TestImageReplayer.NextTag failure
Flagging as resolved by ticket #16708 Jason Dillaman

08/05/2016

02:12 PM Bug #15947 (In Progress): Sporadic TestImageReplayer.NextTag failure
Jason Dillaman
12:36 PM Bug #16921: rbd-nbd IO hang
[ 3799.647869] nbd: registered device at major 43
[ 4068.478821] block nbd0: NBD_DISCONNECT
[ 4068.478895] block nb...
Jason Dillaman
01:01 AM Bug #16887 (Fix Under Review): ceph 10.2.2 rbd status on image format 2 returns "(2) No such file...
*PR*: https://github.com/ceph/ceph/pull/10581
Won't solve the issue for existing images since the API uses fixed w...
Jason Dillaman
01:00 AM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
This is a very old issue that could hit if the combination of a client's (global) instance id concatenated with a pot... Jason Dillaman
12:51 AM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
My assertions of this being jewel specific should be treated with a grain of salt, it may well be that we tipped over... Kjetil Joergensen
12:24 AM Bug #16887 (In Progress): ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or d...
Jason Dillaman
12:21 AM Bug #16889 (Need More Info): Ceph 10.2.2 meet a Segmentation fault after rename a image with form...
@de Ian: can you install the debuginfo packages so that the backtrace can resolve the full call stack? It quite possi... Jason Dillaman

08/04/2016

08:27 PM Bug #16921 (New): rbd-nbd IO hang
Running fsx_nbd on my local trusty VM works just fine. Examining the test logs, they appear to be missing the first c... Jason Dillaman
06:13 PM Bug #16921 (In Progress): rbd-nbd IO hang
Jason Dillaman
02:45 PM Bug #16921 (Resolved): rbd-nbd IO hang
It can be reproduced on jewel with *rbd/thrash/{base/install.yaml clusters/{fixed-2.yaml openstack.yaml} fs/xfs.yaml ... Loïc Dachary
06:11 PM Bug #16923 (Fix Under Review): librbd/ExclusiveLock.cc: 197: FAILED assert(m_watch_handle != 0)
*PR*: https://github.com/ceph/ceph/pull/10574 Jason Dillaman
04:47 PM Bug #16923 (Resolved): librbd/ExclusiveLock.cc: 197: FAILED assert(m_watch_handle != 0)
ceph version 10.2.0 (3a9fba20ec743699b69bd0181dd6c54dc01c64b9)
1: (()+0x24c4bb) [0x7fdebcbde4bb]
2: (()+0x...
Jason Dillaman

08/03/2016

05:45 AM Backport #16904 (Resolved): jewel: journal should prefetch small chunks of the object during replay
https://github.com/ceph/ceph/pull/10684 Nathan Cutler
05:44 AM Backport #16903 (Resolved): jewel: Non-primary image is recording journal events during image sync
https://github.com/ceph/ceph/pull/10797 Nathan Cutler
05:44 AM Backport #16902 (Resolved): jewel: rbd-mirror: image deleter should use pool id + global image uu...
https://github.com/ceph/ceph/pull/11433 Nathan Cutler
03:25 AM Bug #16898 (Resolved): "TestLibRBD.UpdateFeatures" tests failed in upgrade:client-upgrade-inferna...
run http://pulpito.ceph.com/kchai-2016-08-02_07:44:12-upgrade-wip-16507-jewel---basic-mira/
jobs: 346861
logs: http...
Kefu Chai

08/02/2016

07:16 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
Patch is for illustrative purposes - not intended as a solution. Kjetil Joergensen
07:15 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
Less broken formatting:... Kjetil Joergensen
07:14 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
To un-break my previous comment slightly:
librbd's create_v2 makes the block prefix / id as: bid_ss << std::hex <<...
Kjetil Joergensen
05:45 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
Actually - if I read this correctly - somehow we consistently end up with block_prefix (including rbd_data.) that's 2... Kjetil Joergensen
01:59 AM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
And - it's not consistent. I have one 10.2.2 cluster where this is a problem and one where it's not. Is the length of... Kjetil Joergensen
05:52 PM Bug #16227 (Resolved): rbd-mirror: volume rename followed by delete, not deleted on secondary
Mykola Golub
05:51 PM Bug #16538 (Pending Backport): rbd-mirror: image deleter should use pool id + global image uuid f...
Mykola Golub
05:50 PM Bug #16478 (Pending Backport): Non-primary image is recording journal events during image sync
Mykola Golub
10:48 AM Bug #16889: Ceph 10.2.2 meet a Segmentation fault after rename a image with format 1
de lan wrote:
> my test result of v10.2.2:http://www.daisycloud.org:9091/teuthology-2016-08-02_16:01:23-rbd:cli-v10....
de lan
10:12 AM Bug #16889: Ceph 10.2.2 meet a Segmentation fault after rename a image with format 1
my test result of v10.2.2:http://www.daisycloud.org:9091/teuthology-2016-08-02_16:01:23-rbd:cli-v10.2.2---basic-plana... de lan
09:07 AM Bug #16889 (Can't reproduce): Ceph 10.2.2 meet a Segmentation fault after rename a image with for...
Hi!,when i test zhe ceph 10.2.2,it often meet a Segmentation fault after rename a image with format 1。
The CI test...
de lan
06:30 AM Bug #16855: rbd mirror: after promote, the mirror image often be up+error
@Jason Dillaman
Hi.
I have repoduced it,and taken zhe log.
it shows that zhe image is split-brained,and it didn't...
de lan

08/01/2016

11:54 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
Summary: It looks like the block_name_prefix/rbd_id may accidentally have been extended by one byte, "rbd status" doe... Kjetil Joergensen
11:35 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
For anybody else with the same problem, really-dirty-hack full of awful assumptions: rados -p rbd listwatchers rbd_he... Kjetil Joergensen
09:58 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
As for - why this impacts us. We use/abuse a combination of advisory locking (hint) and watchers to paper over mandat... Kjetil Joergensen
08:43 PM Bug #16887: ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or directory"
With --debug-rados=20/20
Ok - so there seems to be some disagreement about what the rbd_header. object should be n...
Kjetil Joergensen
08:05 PM Bug #16887 (Resolved): ceph 10.2.2 rbd status on image format 2 returns "(2) No such file or dire...
Not quite sure when this started failing, it only ended up noticeable now. So, I can't tell you exactly when, but I'm... Kjetil Joergensen

07/30/2016

04:34 PM Bug #16223 (Pending Backport): journal should prefetch small chunks of the object during replay
Mykola Golub
03:25 AM Backport #16869 (Resolved): jewel: Discard hangs when 'rbd_skip_partial_discard' is enabled
https://github.com/ceph/ceph/pull/10797 Nathan Cutler
03:25 AM Backport #16868 (Resolved): jewel: Prevent the creation of a clone from a non-primary mirrored image
https://github.com/ceph/ceph/pull/10650 Nathan Cutler
03:24 AM Backport #16867 (Resolved): jewel: mkfs.xfs slow performance with discards and object map
https://github.com/ceph/ceph/pull/10649 Nathan Cutler

07/29/2016

03:56 PM Bug #15353 (Rejected): librbd: disable optimizations that result in pipelining guarded writes mix...
Yes, now that we store write errors in the pg log this shouldn't be an issue. Josh Durgin
03:04 PM Bug #15353 (Need More Info): librbd: disable optimizations that result in pipelining guarded writ...
@Josh: do the recent PG log changes make this ticket obsolete? Jason Dillaman
03:07 PM Bug #15871 (Resolved): Replay of snap remove journal event caused assertion failure
Fixed under ticket #16114 and was backported via same ticket. Jason Dillaman
03:05 PM Bug #15561 (Closed): RBD image can be listed but not opened
@Ke Ke: feel free to re-open if you are still experiencing this issue. Jason Dillaman
02:55 PM Bug #16740 (In Progress): Cannot disable journaling or remove non-mirrored, "non-primary" image
Jason Dillaman
01:31 PM Bug #16689 (Pending Backport): mkfs.xfs slow performance with discards and object map
Mykola Golub
12:34 PM Bug #16855 (Need More Info): rbd mirror: after promote, the mirror image often be up+error
@de Ian: can you please provide a debug log from the rbd-mirror daemon executing against cluster 'cluster1'? You need... Jason Dillaman
07:28 AM Bug #16855 (Resolved): rbd mirror: after promote, the mirror image often be up+error
HI!
when i do some demote and promote operation.zhe mirror image often become up+error....
de lan
04:28 AM Feature #13186: I hope retain snapshot of the rbd block, after rbd block export and import
zouming zou wrote:
> (1)I have a rbd block foo_c01,that contain a snapshot foo_c01_s01.then,I export foo_c01 to foo_...
Yang Dongsheng

07/28/2016

08:46 PM Bug #16227 (Fix Under Review): rbd-mirror: volume rename followed by delete, not deleted on secon...
*PR*: https://github.com/ceph/ceph/pull/10484
(will be backported with ticket #16538)
Jason Dillaman
08:25 PM Bug #16227 (In Progress): rbd-mirror: volume rename followed by delete, not deleted on secondary
Jason Dillaman
08:46 PM Bug #16538 (Fix Under Review): rbd-mirror: image deleter should use pool id + global image uuid f...
*PR*: https://github.com/ceph/ceph/pull/10484 Jason Dillaman
07:19 PM Feature #16171 (Fix Under Review): Request exclusive lock if owner sends -ENOTSUPP for proxied ma...
*PR*: https://github.com/ceph/ceph/pull/10481 Jason Dillaman
03:38 PM Feature #16171 (In Progress): Request exclusive lock if owner sends -ENOTSUPP for proxied mainten...
Jason Dillaman
06:04 PM Bug #16386 (Pending Backport): Discard hangs when 'rbd_skip_partial_discard' is enabled
Jason Dillaman

07/27/2016

12:23 PM Feature #15388 (Resolved): Inherit the Parent Image properties while Cloning a rbd Image
Jason Dillaman
12:22 PM Feature #6626 (Resolved): openstack: cinder: allow users to delete snapshots that have clones
Jason Dillaman

07/26/2016

02:56 PM Bug #16717: "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrade-jewel-distro...
run: http://pulpito.ceph.com/teuthology-2016-07-20_02:45:02-upgrade:client-upgrade-jewel-distro-basic-smithi/
Jobs: ...
Yuri Weinstein

07/25/2016

06:35 PM Bug #16811 (New): [udev] /dev/rbd/<poolname>/<imagename> symlink should have an fsid in it
It's possible to map images from different clusters on the same box. Currently, if (poolname, imagename) pair happen... Ilya Dryomov
05:03 PM Bug #16478 (In Progress): Non-primary image is recording journal events during image sync
Jason Dillaman
05:02 PM Bug #16741: io getting stuck after advancing journal object set
@Mykola: do you think this is still and issue after the recent changes introduced in the reduce memory footprint branch? Jason Dillaman
04:50 PM Bug #16708 (Fix Under Review): Sporadic failure in TestImageReplayer.StartReplayAndWrite
*PR*: https://github.com/ceph/ceph/pull/10432 Jason Dillaman
05:58 AM Bug #16394: Ceph RBD
junming rao wrote:
> @Jason
> This is debug logs from librbd;
> thanks。
junming rao
05:56 AM Bug #16394: Ceph RBD
@Jason
This is debug logs from librbd;
thanks。
junming rao

07/24/2016

03:13 PM Cleanup #16130 (Resolved): Proxied operations shouldn't result in error messages if replayed
Jason Dillaman
03:13 PM Bug #16449 (Pending Backport): Prevent the creation of a clone from a non-primary mirrored image
Jason Dillaman
03:10 PM Bug #16717 (In Progress): "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrad...
@Mykola: instead of waiting (possibly for a very long time) for the backport, perhaps just change the test [1] to ove... Jason Dillaman

07/23/2016

09:47 PM Bug #16799 (Duplicate): hammer-backports: rbd import fails in cache tier test
Nathan Cutler
08:17 PM Bug #16799 (Duplicate): hammer-backports: rbd import fails in cache tier test
test: rados/singleton-nomsgr/{all/export-after-evict.yaml}
what happens:...
Nathan Cutler

07/22/2016

10:17 PM Bug #16708: Sporadic failure in TestImageReplayer.StartReplayAndWrite
also in https://jenkins.ceph.com/job/ceph-pull-requests/9476/consoleFull#-1053578855d63714d2-c8d8-41fc-a9d4-8dee30be4c32 Josh Durgin
01:10 PM Bug #16708 (In Progress): Sporadic failure in TestImageReplayer.StartReplayAndWrite
Jason Dillaman
06:21 AM Bug #16708: Sporadic failure in TestImageReplayer.StartReplayAndWrite
spotted again in https://jenkins.ceph.com/job/ceph-pull-requests/9474/consoleFull#-1053578855d63714d2-c8d8-41fc-a9d4-... Kefu Chai
08:37 PM Backport #16796 (Resolved): jewel: Renaming old format image results in "Transport endpoint is no...
https://github.com/ceph/ceph/pull/10684 Nathan Cutler
04:30 PM Feature #16780 (Resolved): rbd-mirror: use sparse read during image sync
If 1 byte is used in a primary image backing object, the image sync process will read and write a full object size ch... Jason Dillaman
07:03 AM Bug #16773 (Resolved): FAILED assert(m_image_ctx.journal == nullptr)
Hi!
when i test the ci suite:rbd:valgrind/{base/install.yaml clusters/{fixed-1.yaml openstack.yaml} fs/xfs.yaml va...
de lan

07/21/2016

09:15 PM Bug #16529: "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-distro-basic-vps
http://qa-proxy.ceph.com/teuthology/teuthology-2016-07-20_04:20:03-upgrade:jewel-x-master-distro-basic-vps/325448/teu... Yuri Weinstein
06:56 PM Bug #16223 (Fix Under Review): journal should prefetch small chunks of the object during replay
Jason Dillaman
03:48 PM Bug #15947: Sporadic TestImageReplayer.NextTag failure
Another instance: https://jenkins.ceph.com/job/ceph-pull-requests/9411/console... Josh Durgin
10:51 AM Bug #16555 (In Progress): librbd should permit removal of image being bootstrapped by rbd-mirror
Mykola Golub
08:05 AM Feature #15632 (Fix Under Review): Expose librbd API methods to directly acquire and release the ...
PR: https://github.com/ceph/ceph/pull/9592 Mykola Golub
08:04 AM Bug #16321 (Pending Backport): Renaming old format image results in "Transport endpoint is not co...
Mykola Golub
07:58 AM Feature #14738 (Fix Under Review): Optionally unregister "laggy" journal clients
PR: https://github.com/ceph/ceph/pull/10378 Mykola Golub

07/20/2016

05:16 AM Backport #16747 (Resolved): jewel: rbd-mirror: snap rename does not correctly replicate
https://github.com/ceph/ceph/pull/10684 Nathan Cutler

07/19/2016

06:49 PM Bug #16741 (Resolved): io getting stuck after advancing journal object set
I can easily reproduce this issue running write bench on an image that has small object size journal:... Mykola Golub
05:19 PM Bug #16622 (Pending Backport): rbd-mirror: snap rename does not correctly replicate
Mykola Golub
04:22 PM Bug #16740 (Resolved): Cannot disable journaling or remove non-mirrored, "non-primary" image
Use rbd-mirror to create a non-primary image. If the 'rbd_mirroring' object is removed, this image will now be treate... Jason Dillaman
12:24 PM Bug #16529: "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-distro-basic-vps
The upgrade test installs 10.2.0, upgrades the OSDs to a point release but keeps the test_cls_rbd at 10.2.0. There wa... Jason Dillaman
11:23 AM Bug #16707 (Fix Under Review): rbd-replay-prep doesn't record discard IO events
*PR*: https://github.com/ceph/ceph/pull/10332 Jason Dillaman
09:17 AM Bug #16717 (Fix Under Review): "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-u...
PR: https://github.com/ceph/ceph/pull/10348 Mykola Golub
07:37 AM Bug #16717 (In Progress): "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrad...
Mykola Golub
07:35 AM Bug #16717: "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrade-jewel-distro...
I guess running hammer ceph_test_librbd_api with jewel librbd is intentional (to run only hammer tests using jewel li... Mykola Golub
07:27 AM Bug #16717: "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrade-jewel-distro...
The teuthology suit upgrades client host from hammer to jewel and reverts ceph_test_librbd_api to hammer using this p... Mykola Golub
07:27 AM Backport #16735 (Resolved): jewel: rbd-nbd does not properly handle resize notifications
https://github.com/ceph/ceph/pull/10679 Nathan Cutler
05:15 AM Bug #16223: journal should prefetch small chunks of the object during replay
*PR*: https://github.com/ceph/ceph/pull/10341 Jason Dillaman
05:07 AM Bug #15715 (Pending Backport): rbd-nbd does not properly handle resize notifications
Mykola Golub

07/18/2016

06:07 PM Bug #16689 (Fix Under Review): mkfs.xfs slow performance with discards and object map
*PR*: https://github.com/ceph/ceph/pull/10332 Jason Dillaman
05:37 PM Bug #16689 (In Progress): mkfs.xfs slow performance with discards and object map
Jason Dillaman
02:38 PM Bug #16689: mkfs.xfs slow performance with discards and object map
Attaching the raw lttng files for librbd runs. The new image is 10G instead of the 100G in the previous case, but all... Vaibhav Bhembre
04:51 PM Bug #16717 (Resolved): "[ FAILED ] TestLibRBD.TestCreateLsDeletePP" in upgrade:client-upgrade-j...
Run: http://pulpito.ceph.com/teuthology-2016-07-13_02:45:02-upgrade:client-upgrade-jewel-distro-basic-smithi/
Job: 3...
Yuri Weinstein
12:26 PM Bug #16708 (Resolved): Sporadic failure in TestImageReplayer.StartReplayAndWrite
From Mykola:... Jason Dillaman
12:22 PM Bug #16223 (In Progress): journal should prefetch small chunks of the object during replay
Jason Dillaman
11:53 AM Bug #16707 (Resolved): rbd-replay-prep doesn't record discard IO events
Jason Dillaman
07:07 AM Bug #16529: "[ FAILED ] TestClsRbd.mirror_image" in upgrade:jewel-x-master-distro-basic-vps
http://pulpito.ceph.com/teuthology-2016-07-17_04:20:03-upgrade:jewel-x-master-distro-basic-vps/319486/
http://pulpit...
Loïc Dachary
 

Also available in: Atom