Activity
From 02/27/2017 to 03/28/2017
03/28/2017
- 04:39 PM Bug #19400 (Resolved): add more info during pool delete error
- In luminous the mon_allow_pool_delete is default to false and it may be confusing for any admin who
tries to delete ... - 02:55 PM Bug #18924: kraken-bluestore 11.2.0 memory leak issue
We decided to stop the ceph-mgr service in all the nodes because is using lot of CPU and we understood that this se...- 02:51 PM Bug #18924: kraken-bluestore 11.2.0 memory leak issue
Fixed with the following commands:
The memory is released by applying the following commands in a content no...- 09:40 AM Documentation #18986: Need to document monitor health configuration values
- the description of "mon warn osd usage percent" and "mon_osd_min_in_ratio " can be found at https://github.com/ceph/...
03/27/2017
- 09:09 PM Bug #19267: rados list-inconsistent-obj sometimes doesn't flag that all 3 copies are bad
- Oh I see, it's missing the error string.
I'm not sure if in this case it's just taking one of them as authoritativ... - 08:59 AM Bug #19320: Pg inconsistent make ceph osd down
- backtrace in the attached log_inconsistent.txt...
03/24/2017
- 10:01 PM Feature #19384 (New): ceph_objectstore_tool (set|clear)-missing-item command
- This one is only relevant for kraken and later. It would be good to have a command for directly manipulating a pg's ...
- 09:59 PM Feature #19383 (New): ceph_objectstore_tool: set-version op to allow setting the prior_version an...
- The motivation for this one is to be able to manually do part of what mark_unfound_lost revert does automatically and...
- 09:49 PM Bug #19380 (New): only sort of a bug: it's possible to get an unfound object without losing min_s...
- Fundamentally, ReplicatedBackend does destructive updates. That makes the following sequence possible. Assume that ...
- 09:33 PM Bug #19379 (Resolved): bluestore: crc mismatch after recent overwrite
- ...
- 08:37 PM Bug #19377: mark_unfound_lost revert won't actually recover the objects unless there are some fou...
- There is a very clumsy workaround to this issue. Once the mark_unfound_lost revert commands claims to have completed...
- 08:34 PM Bug #19377 (Duplicate): mark_unfound_lost revert won't actually recover the objects unless there ...
- See ReplicatedPG::start_recovery_ops. If the num_missing==num_unfound, we don't try to do recovery. This is problem...
03/23/2017
- 01:36 PM Bug #18698: BlueFS FAILED assert(0 == "allocate failed... wtf")
- Hi !
I also got that issue. I also added "bluefs_allocator = stupid" in my /etc/ceph/ceph.conf. Worked.
Issue happe...
03/22/2017
- 10:14 PM Bug #18162: osd/ReplicatedPG.cc: recover_replicas: object added to missing set for backfill, but ...
- I've created https://github.com/ceph/ceph/pull/14054 to track Alexandre's changes.
I'm working on handling out of ... - 09:10 AM Bug #19348 (Can't reproduce): "ceph ping mon.c" cli prints assertion failure on timeout
- # start a cluster with 3 monitors: mon.a, mon.b and mon.c
# stop mon.c
# ceph ping mon.c --connect-timeout=5
it ...
03/21/2017
- 10:09 AM Bug #19320 (New): Pg inconsistent make ceph osd down
- Hi all.
I am running a ceph cluster.
These is a pg inconsistent:
pg 3.aff is active+recovery_wait+degraded+incon...
03/20/2017
- 04:26 AM Feature #15835: filestore: randomize split threshold
- Hi! I am an undergrad student wishing to contribute to CEPH, and I would like to work on this issue. Please let me kn...
03/18/2017
- 06:03 AM Bug #19267: rados list-inconsistent-obj sometimes doesn't flag that all 3 copies are bad
- Greg Farnum wrote:
> I don't understand. What about this output says that two copies are bad and one isn't?
Thank...
03/17/2017
- 09:33 PM Bug #19300 (Can't reproduce): "Segmentation fault ceph_test_objectstore --gtest_filter=-*/3"
- Run: http://pulpito.ceph.com/yuriw-2017-03-16_15:10:12-rados-wip-yuri-testing_2017_3_16-distro-basic-smithi/
Logs: h... - 09:11 PM Bug #19267: rados list-inconsistent-obj sometimes doesn't flag that all 3 copies are bad
- I don't understand. What about this output says that two copies are bad and one isn't?
- 08:11 PM Bug #19299: Jewel -> Kraken: OSD boot takes 1+ hours, unusually high CPU
- As far as I know that was with -f passed. Maybe is relevant that I grepped out 'madvise' calls because they occur at...
- 07:59 PM Bug #19299 (Need More Info): Jewel -> Kraken: OSD boot takes 1+ hours, unusually high CPU
- 07:58 PM Bug #19299: Jewel -> Kraken: OSD boot takes 1+ hours, unusually high CPU
- the strace doesn't include child processes.. can you repeat with -f passed to strace?
- 07:48 PM Bug #19299 (Can't reproduce): Jewel -> Kraken: OSD boot takes 1+ hours, unusually high CPU
- Since upgrading to Kraken we've had severe problems with OSD startup. Though this ticket mentions bootup specificall...
03/13/2017
- 09:48 AM Bug #19267 (New): rados list-inconsistent-obj sometimes doesn't flag that all 3 copies are bad
- I tested ceph 10.2.3 with cluster of 3 osd nodes.
I upload a text file to ceph cluster, then manually change the t...
03/09/2017
- 11:55 PM Bug #18926: Why osds do not release memory?
I don't know if this is the same issue, but bluestore got EROFS (30) and asserted.
http://pulpito.ceph.com/dzaf...- 03:54 PM Bug #18687 (Resolved): bluestore: ENOSPC writing to XFS block file on smithi
- Fixed as part of https://github.com/ceph/ceph/pull/13425 for tracker 15912
03/08/2017
- 09:38 PM Bug #19237 (New): "PG.cc: 3100: FAILED assert(e.version > info.last_update)" in upgrade:kraken-x-...
- Run: http://pulpito.ceph.com/teuthology-2017-03-08_02:25:22-upgrade:kraken-x-master-distro-basic-vps/
Job: 894399
L... - 12:17 PM Feature #18943 (Resolved): crush: add devices class that rules can use as a filter
03/07/2017
- 07:24 AM Bug #19191: osd/ReplicatedBackend.cc: 1109: FAILED assert(!parent->get_log().get_missing().is_mis...
- Do we actually want to clear the missing set here, or just filter it for the correct child PG?
...I presume killing ... - 07:19 AM Bug #19191: osd/ReplicatedBackend.cc: 1109: FAILED assert(!parent->get_log().get_missing().is_mis...
- Summary:
During pg split we are resetting last_backfill but not clearing the
local missing set. This comes back t... - 01:19 AM Bug #19199: Odd OSD failure path; ERROR: osd init failed: (110) Connection timed out
- Earlier in the log the root cause appears:...
03/06/2017
- 06:48 PM Bug #19199 (New): Odd OSD failure path; ERROR: osd init failed: (110) Connection timed out
- See attached OSD log for more details.
commit 6f8e4b38103d6f519e6661acc97a47ceccf5e5fc was the latest master
Interm... - 05:05 PM Bug #19198 (Closed): Bluestore doubles mem usage when caching object content
- When trying to cache object content BlueStore uses twice as much memory than it really caches.
The root cause for ... - 09:48 AM Bug #19191: osd/ReplicatedBackend.cc: 1109: FAILED assert(!parent->get_log().get_missing().is_mis...
- yuri, the backtrace you posted is another issue. i am building your branch of wip-yuri-testing_2017_3_4 to see if "ce...
- 01:22 AM Bug #19191: osd/ReplicatedBackend.cc: 1109: FAILED assert(!parent->get_log().get_missing().is_mis...
- Also see during PRs testing https://trello.com/c/il60a5yB
http://qa-proxy.ceph.com/teuthology/yuriw-2017-03-05_23:...
03/04/2017
- 07:24 PM Bug #19191 (Resolved): osd/ReplicatedBackend.cc: 1109: FAILED assert(!parent->get_log().get_missi...
- ...
02/27/2017
- 10:03 AM Bug #19092 (New): cluster [ERR] scrub 2.1 ... is an unexpected clone" in cluster log
- see http://pulpito.ceph.com/kchai-2017-02-27_04:13:29-rados-wip-kefu-testing---basic-smithi/862801/
after evicting... - 06:53 AM Feature #18943: crush: add devices class that rules can use as a filter
- https://github.com/ceph/ceph/pull/13444
- 03:27 AM Bug #19086 (Need More Info): BlockDevice::create should add check for readlink result instead of ...
Also available in: Atom