Project

General

Profile

Activity

From 09/26/2017 to 10/25/2017

10/25/2017

11:38 PM Bug #21568 (Pending Backport): MDSMonitor commands crashing on cluster upgraded from Hammer (none...
Patrick Donnelly
11:37 PM Bug #21821 (Pending Backport): MDSMonitor: mons should reject misconfigured mds_blacklist_interval
Patrick Donnelly
11:36 PM Bug #20596 (Resolved): MDSMonitor: obsolete `mds dump` and other deprecated mds commands
Patrick Donnelly
08:37 PM Bug #21843 (Pending Backport): mds: preserve order of requests during recovery of multimds cluster
Patrick Donnelly
08:35 PM Bug #21848 (In Progress): client: re-expand admin_socket metavariables in child process
Zhi, please revisit this issue as the fix in https://github.com/ceph/ceph/pull/18393 must be reverted due to the reas... Patrick Donnelly
06:14 PM Bug #21928 (Resolved): src/mds/MDCache.cc: 7421: FAILED assert(CInode::count() == inode_map.size(...
... Patrick Donnelly
02:55 PM Feature #21888: Adding [--repair] option for cephfs-journal-tool make it can recover all journal ...
Assigning to Ivan, as he had already submitted a PR for this Feature. Jos Collin
01:50 PM Bug #18743 (Fix Under Review): Scrub considers dirty backtraces to be damaged, puts in damage tab...
https://github.com/ceph/ceph/pull/18538
Inspired to fix this from working on today's "[ceph-users] MDS damaged" th...
John Spray
07:53 AM Bug #21393 (In Progress): MDSMonitor: inconsistent role/who usage in command help
Jos Collin
06:43 AM Bug #21903: ceph-volume-client: File "cephfs.pyx", line 696, in cephfs.LibCephFS.open
Patrick Donnelly wrote:
> Another one during retest: http://pulpito.ceph.com/pdonnell-2017-10-24_16:24:48-fs-wip-pdo...
Ramana Raja

10/24/2017

09:15 PM Bug #21908: kcephfs: mount fails with -EIO
... Patrick Donnelly
09:14 PM Bug #21908 (New): kcephfs: mount fails with -EIO
... Patrick Donnelly
08:23 PM Bug #21884: client: populate f_fsid in statfs output
Jeff, I liked your suggestion of a vxattr the client can lookup to check if the mount is CephFS. Patrick Donnelly
05:17 PM Bug #21903: ceph-volume-client: File "cephfs.pyx", line 696, in cephfs.LibCephFS.open
Another one during retest: http://pulpito.ceph.com/pdonnell-2017-10-24_16:24:48-fs-wip-pdonnell-testing-201710240410-... Patrick Donnelly
05:17 PM Bug #21903 (New): ceph-volume-client: File "cephfs.pyx", line 696, in cephfs.LibCephFS.open
... Patrick Donnelly
03:24 PM Bug #21393 (Fix Under Review): MDSMonitor: inconsistent role/who usage in command help
https://github.com/ceph/ceph/pull/18512 Jos Collin

10/23/2017

07:51 PM Bug #21884: client: populate f_fsid in statfs output
I'm not sure we want to change f_type. It _is_ still FUSE, regardless of what userland daemon it's talking to.
If ...
Jeff Layton
10:43 AM Bug #21884: client: populate f_fsid in statfs output
Here's how the kernel client fills this field out:... Jeff Layton
11:46 AM Bug #21848 (Fix Under Review): client: re-expand admin_socket metavariables in child process
Nathan Cutler
02:28 AM Bug #21393 (In Progress): MDSMonitor: inconsistent role/who usage in command help
Jos Collin
01:51 AM Bug #21892 (Resolved): limit size of subtree migration
... Zheng Yan

10/22/2017

09:19 AM Feature #21888: Adding [--repair] option for cephfs-journal-tool make it can recover all journal ...
https://github.com/ceph/ceph/pull/18465/ Ivan Guan
06:15 AM Feature #21888 (Fix Under Review): Adding [--repair] option for cephfs-journal-tool make it can r...
As described in the document if a journal is damaged or for any reason an MDS is incapable of replaying it, attempt t... Ivan Guan

10/21/2017

04:05 AM Bug #21884 (Resolved): client: populate f_fsid in statfs output
We should just reuse the kclient -f_id- f_type as, in principle, the application should only need to know that the fi... Patrick Donnelly

10/20/2017

03:21 PM Bug #21512: qa: libcephfs_interface_tests: shutdown race failures
PR to master was https://github.com/ceph/ceph/pull/18139 Ken Dreyer
10:08 AM Feature #21877 (Resolved): quota and snaprealm integation
https://github.com/ceph/ceph/pull/18424/commits/4477f8b93d183eb461798b5b67550d3d5b22c16c Zheng Yan
09:30 AM Backport #21874 (Resolved): luminous: qa: libcephfs_interface_tests: shutdown race failures
https://github.com/ceph/ceph/pull/20082 Nathan Cutler
09:29 AM Backport #21870 (Resolved): luminous: Assertion in EImportStart::replay should be a damaged()
https://github.com/ceph/ceph/pull/18930 Nathan Cutler
07:10 AM Bug #21861 (New): osdc: truncate Object and remove the bh which have someone wait for read on it ...
ceph version: jewel 10.2.2
When one osd be written over the full_ratio(default is 0.95) will lead the cluster t...
Ivan Guan

10/19/2017

11:22 PM Bug #21777 (Need More Info): src/mds/MDCache.cc: 4332: FAILED assert(mds->is_rejoin())
This is NMI because we weren't able to reproduce the actual problem. We'll ahve to wait for QE to reproduce again wit... Patrick Donnelly
06:35 PM Bug #21853 (Resolved): mds: mdsload debug too high
... Patrick Donnelly
01:44 PM Feature #19578: mds: optimize CDir::_omap_commit() and CDir::_committed() for large directory
this should help large directory performance Zheng Yan
07:52 AM Bug #21848: client: re-expand admin_socket metavariables in child process
https://github.com/ceph/ceph/pull/18393 Zhi Zhang
07:51 AM Bug #21848 (Resolved): client: re-expand admin_socket metavariables in child process
The default value of admin_socket is $run_dir/$cluster-$name.asok. If mounting multiple ceph-fuse instances on the sa... Zhi Zhang
07:44 AM Bug #21483 (Resolved): qa: test_snapshot_remove (kcephfs): RuntimeError: Bad data at offset 0
Zheng Yan
02:20 AM Bug #21749 (Duplicate): PurgeQueue corruption in 12.2.1
dup of #19593 Zheng Yan
02:17 AM Backport #21658 (Fix Under Review): luminous: purge queue and standby replay mds
https://github.com/ceph/ceph/pull/18385 Zheng Yan
01:53 AM Bug #21843 (Fix Under Review): mds: preserve order of requests during recovery of multimds cluster
https://github.com/ceph/ceph/pull/18384 Zheng Yan
01:50 AM Bug #21843 (Resolved): mds: preserve order of requests during recovery of multimds cluster
there are several cases that requests get processed in wrong order
1)
touch a/b/f (handled by mds.1, early ...
Zheng Yan

10/17/2017

10:09 PM Bug #21821 (Fix Under Review): MDSMonitor: mons should reject misconfigured mds_blacklist_interval
https://github.com/ceph/ceph/pull/18366 John Spray
09:55 PM Bug #21821: MDSMonitor: mons should reject misconfigured mds_blacklist_interval
A good opportunity to use the new min/max fields on the config option itself.
I suppose if we accept the idea that...
John Spray
05:36 PM Bug #21821 (Resolved): MDSMonitor: mons should reject misconfigured mds_blacklist_interval
There should be a minimum acceptable value otherwise we see potential behavior where a blacklisted MDS is still writi... Patrick Donnelly
02:01 AM Bug #21812 (Closed): standby replay mds may submit log
I wrongly interpret the log Zheng Yan

10/16/2017

01:19 PM Bug #21812 (Closed): standby replay mds may submit log
magna002://home/smohan/LOGS/cfs-mds.magna116.trunc.log.gz
mds submitted log entry while it's in standby replay sta...
Zheng Yan
12:07 AM Bug #21807 (Pending Backport): mds: trims all unpinned dentries when memory limit is reached
Patrick Donnelly
12:03 AM Backport #21810 (Resolved): luminous: mds: trims all unpinned dentries when memory limit is reached
https://github.com/ceph/ceph/pull/18316 Patrick Donnelly

10/14/2017

08:49 PM Bug #21807 (Fix Under Review): mds: trims all unpinned dentries when memory limit is reached
https://github.com/ceph/ceph/pull/18309 Patrick Donnelly
08:46 PM Bug #21807 (Resolved): mds: trims all unpinned dentries when memory limit is reached
Generally dentries are pinned by the client cache so this was easy to miss in testing. Bug is here:
https://github...
Patrick Donnelly
08:17 AM Backport #21805 (In Progress): luminous: client_metadata can be missing
Nathan Cutler
12:32 AM Backport #21805 (Resolved): luminous: client_metadata can be missing
https://github.com/ceph/ceph/pull/18299 Patrick Donnelly
08:16 AM Backport #21804 (In Progress): luminous: limit internal memory usage of object cacher.
Nathan Cutler
12:23 AM Backport #21804 (Resolved): luminous: limit internal memory usage of object cacher.
https://github.com/ceph/ceph/pull/18298 Patrick Donnelly
12:39 AM Backport #21806 (In Progress): luminous: FAILED assert(in->is_dir()) in MDBalancer::handle_export...
Patrick Donnelly
12:37 AM Backport #21806 (Resolved): luminous: FAILED assert(in->is_dir()) in MDBalancer::handle_export_pi...
https://github.com/ceph/ceph/pull/18300 Patrick Donnelly
12:16 AM Bug #21512 (Pending Backport): qa: libcephfs_interface_tests: shutdown race failures
Patrick Donnelly
12:14 AM Bug #21726 (Pending Backport): limit internal memory usage of object cacher.
Patrick Donnelly
12:14 AM Bug #21746 (Pending Backport): client_metadata can be missing
Patrick Donnelly
12:13 AM Bug #21759 (Pending Backport): Assertion in EImportStart::replay should be a damaged()
Patrick Donnelly
12:12 AM Bug #21768 (Pending Backport): FAILED assert(in->is_dir()) in MDBalancer::handle_export_pins()
Patrick Donnelly

10/13/2017

07:18 PM Feature #21601 (Resolved): ceph_volume_client: add get, put, and delete object interfaces
Patrick Donnelly
07:18 PM Backport #21602 (Resolved): luminous: ceph_volume_client: add get, put, and delete object interfaces
Patrick Donnelly
12:37 AM Bug #21777 (Fix Under Review): src/mds/MDCache.cc: 4332: FAILED assert(mds->is_rejoin())
https://github.com/ceph/ceph/pull/18278 Patrick Donnelly

10/12/2017

10:47 PM Bug #21777 (Need More Info): src/mds/MDCache.cc: 4332: FAILED assert(mds->is_rejoin())
MDS may send a MMDSCacheRejoin(MMDSCacheRejoin::OP_WEAK) message to an MDS which is not rejoin/active/stopping. Once ... Patrick Donnelly
04:04 AM Bug #21768 (Fix Under Review): FAILED assert(in->is_dir()) in MDBalancer::handle_export_pins()
https://github.com/ceph/ceph/pull/18261 Zheng Yan
03:58 AM Bug #21768 (Resolved): FAILED assert(in->is_dir()) in MDBalancer::handle_export_pins()
... Zheng Yan

10/11/2017

09:38 PM Bug #21765: auth|doc: fs authorize error for existing credentials confusing/unclear
Doug, please take this one. Patrick Donnelly
09:37 PM Bug #21765 (Resolved): auth|doc: fs authorize error for existing credentials confusing/unclear
If you attempt to use `fs authorize` on a key that already exists you get an error like:
https://github.com/ceph/c...
Patrick Donnelly
03:47 PM Bug #21764 (Resolved): common/options.cc: Update descriptions and visibility levels for MDS/clien...
Go through the options in common/options.cc and figure out which should be LEVEL_DEV (hidden in the UI). BASIC/ADVANC... Patrick Donnelly
12:02 PM Bug #21748: client assertions tripped during some workloads
Huh. That is an interesting theory. I don't see how ganesha would do that, but maybe. Unfortunately, the original pro... Jeff Layton
08:27 AM Bug #21748: client assertions tripped during some workloads
this shouldn't happen even for traceless reply. I suspect the 'in' passed to ceph_ll_setattr isn't belong to the 'cmo... Zheng Yan
11:03 AM Bug #21759 (Fix Under Review): Assertion in EImportStart::replay should be a damaged()
https://github.com/ceph/ceph/pull/18244 John Spray
10:35 AM Bug #21759 (Resolved): Assertion in EImportStart::replay should be a damaged()

This is one of a number of assertions that still linger in journal.cc, but since it's been seen in the wild ("[ceph...
John Spray
09:11 AM Bug #21754: mds: src/osdc/Journaler.cc: 402: FAILED assert(!r)
... Zheng Yan
06:31 AM Bug #21749: PurgeQueue corruption in 12.2.1
Hi Yan,
yes, we had 3 MDS running in standby-replay mode (I switched them to standby now).
Thanks for the offer...
Daniel Baumann
02:58 AM Bug #21749: PurgeQueue corruption in 12.2.1
likely caused by http://tracker.ceph.com/issues/19593.
ping 'yanzheng' at ceph@OFTC, I will help you to recover th...
Zheng Yan
02:40 AM Bug #21745: mds: MDBalancer using total (all time) request count in load statistics
although it is simple to add last_timestamp and last_reqcount so that we can get an average TPS, but TPS may fluctuat... Xiaoxi Chen

10/10/2017

10:10 PM Bug #21754 (Rejected): mds: src/osdc/Journaler.cc: 402: FAILED assert(!r)
... Patrick Donnelly
04:09 PM Bug #21749: PurgeQueue corruption in 12.2.1
I saved all information/logs/objects, feel free to ask for any of it and further things.
Regards,
Daniel
Daniel Baumann
12:05 PM Bug #21749 (Duplicate): PurgeQueue corruption in 12.2.1
From "[ceph-users] how to debug (in order to repair) damaged MDS (rank)?"
Log snippet during MDS startup:...
John Spray
02:18 PM Bug #21748: client assertions tripped during some workloads
Actually this is wrong (as Zheng pointed out). The call is made with a zero-length path that starts from the inode on... Jeff Layton
10:44 AM Bug #21748: client assertions tripped during some workloads
The right fix is probably to just remove that assertion. I don't think it's really valid anyway. cephfs turns the ino... Jeff Layton
10:42 AM Bug #21748 (Can't reproduce): client assertions tripped during some workloads
We had a report of some crashes in ganesha here:
https://github.com/nfs-ganesha/nfs-ganesha/issues/215
Dan and ...
Jeff Layton
12:51 PM Bug #21412: cephfs: too many cephfs snapshots chokes the system
The trimsnap states. The rmdir actually completes quickly, but the resulting operations throw the entire cluster int... Wyllys Ingersoll
02:51 AM Bug #21412: cephfs: too many cephfs snapshots chokes the system
what do you mean "it takes almost 24 hours to delete a single snapshot"? 'rmdir .snap/xxx' tooks 24 hours or pgs on ... Zheng Yan
09:55 AM Bug #21746 (Fix Under Review): client_metadata can be missing
https://github.com/ceph/ceph/pull/18215 Zheng Yan
09:52 AM Bug #21746 (Resolved): client_metadata can be missing
session opened by Server::prepare_force_open_sessions() has no client metadata. Zheng Yan
09:47 AM Bug #21745 (Resolved): mds: MDBalancer using total (all time) request count in load statistics
This was pointed out by Xiaoxi Chen
The get_req_rate() function is returning the value of l_mds_request, which is ...
John Spray

10/09/2017

06:25 PM Bug #21405 (Fix Under Review): qa: add EC data pool to testing
https://github.com/ceph/ceph/pull/18192 Sage Weil
05:50 PM Bug #21734 (Duplicate): mount client shows total capacity of cluster but not of a pool
SERVER:
ceph df
GLOBAL:
SIZE AVAIL RAW USED %RAW USED
44637G...
Petr Malkov
01:56 PM Bug #21412: cephfs: too many cephfs snapshots chokes the system
Note, the bug says "10.2.7" but we have since upgraded to 10.2.9 and the same problem exists. Wyllys Ingersoll
01:55 PM Bug #21412: cephfs: too many cephfs snapshots chokes the system
Here is a dump of the cephfs 'dentry_lru' table, in case it is interesting. Wyllys Ingersoll
01:53 PM Bug #21412: cephfs: too many cephfs snapshots chokes the system
Here is data collected from a recent attempt to delete a very old and very large snapshot:
The snapshot extended a...
Wyllys Ingersoll
10:30 AM Bug #21726 (Fix Under Review): limit internal memory usage of object cacher.
https://github.com/ceph/ceph/pull/18183 Zheng Yan
10:22 AM Bug #21726 (Resolved): limit internal memory usage of object cacher.
https://bugzilla.redhat.com/show_bug.cgi?id=1490814
Zheng Yan
07:07 AM Bug #21722: mds: no assertion on inode being purging in find_ino_peers()
https://github.com/ceph/ceph/pull/18174 Zhi Zhang
07:06 AM Bug #21722 (Resolved): mds: no assertion on inode being purging in find_ino_peers()
Recently we hit an assertion on MDS only few times when MDS was very busy.... Zhi Zhang
03:43 AM Bug #20938: CephFS: concurrent access to file from multiple nodes blocks for seconds
kernel of RHEL7 supports FUSE_AUTO_INVAL_DATA. But FUSE_CAP_DONT_MASK was added in libfuse 3.0. Currently no major li... Zheng Yan

10/06/2017

05:46 PM Feature #15066: multifs: Allow filesystems to be assigned RADOS namespace as well as pool for met...
we should default to using a namespace named after the filesystem unless otherwise specified. Douglas Fuller
05:45 PM Feature #21709 (New): ceph fs authorize should detect the correct data namespace
when per-FS data namespaces are enabled, ceph fs authorize should be updated to issue caps for them Douglas Fuller
02:21 PM Bug #21512: qa: libcephfs_interface_tests: shutdown race failures
I have a PR up that seems to fix this, but it may not be what we need. env_to_vec seems like it ought to be reworked ... Jeff Layton

10/05/2017

12:04 PM Bug #21512: qa: libcephfs_interface_tests: shutdown race failures
Looking now to see if we can somehow just fix up lockdep for this. Most of the problems I have seen have seen are fal... Jeff Layton

10/04/2017

06:42 PM Bug #21512: qa: libcephfs_interface_tests: shutdown race failures
I'm now looking for ways to selectively disable lockdep for just this test. So far, I've been unable to do so:
<pr...
Jeff Layton
04:22 PM Bug #21512: qa: libcephfs_interface_tests: shutdown race failures
Patch to make the ShutdownRace test even more thrashy. This has each thread do the setup and teardown in a tight loop... Jeff Layton
02:32 AM Bug #21568 (Fix Under Review): MDSMonitor commands crashing on cluster upgraded from Hammer (none...
https://github.com/ceph/ceph/pull/18109 Patrick Donnelly

10/03/2017

07:41 PM Bug #20938: CephFS: concurrent access to file from multiple nodes blocks for seconds
Ok, I think you're probably right there. Do we need a cmake test to ensure the fuse library defines FUSE_AUTO_INVAL_D... Jeff Layton
02:59 AM Backport #21658 (Resolved): luminous: purge queue and standby replay mds
https://github.com/ceph/ceph/pull/18385 Nathan Cutler
02:58 AM Backport #21657 (Resolved): luminous: StrayManager::truncate is broken
https://github.com/ceph/ceph/pull/18019 Nathan Cutler

10/02/2017

06:29 PM Backport #21626 (In Progress): jewel: ceph_volume_client: sets invalid caps for existing IDs with...
Patrick Donnelly
06:20 PM Backport #21626 (Resolved): jewel: ceph_volume_client: sets invalid caps for existing IDs with no...
https://github.com/ceph/ceph/pull/18084 Patrick Donnelly
06:29 PM Backport #21627 (In Progress): luminous: ceph_volume_client: sets invalid caps for existing IDs w...
Patrick Donnelly
06:25 PM Backport #21627 (Resolved): luminous: ceph_volume_client: sets invalid caps for existing IDs with...
https://github.com/ceph/ceph/pull/18085
-https://github.com/ceph/ceph/pull/18447-
Patrick Donnelly
12:41 PM Bug #21568: MDSMonitor commands crashing on cluster upgraded from Hammer (nonexistent pool?)
Patrick Donnelly
12:40 PM Bug #21568: MDSMonitor commands crashing on cluster upgraded from Hammer (nonexistent pool?)
User confirmed the MDSMap referred to data pools that no longer exist. The fix should check for non-existent pools an... Patrick Donnelly

10/01/2017

06:01 AM Bug #21304: mds v12.2.0 crashing
The following crash still persists with v12.2.1:
2017-10-01 06:07:34.673356 7f1066040700 0 -- 194.249.156.134:680...
Andrej Filipcic
12:46 AM Bug #19593 (Pending Backport): purge queue and standby replay mds
Patrick Donnelly
12:45 AM Bug #21501 (Pending Backport): ceph_volume_client: sets invalid caps for existing IDs with no caps
Patrick Donnelly

09/29/2017

03:12 PM Bug #21604: mds: may recall all client caps (?) because dirty/pinned metadata is not flushed
This sounds like something controlled by config options. The MDS wants to batch up data when flushing to avoid gratui... Greg Farnum
03:02 PM Bug #21604 (New): mds: may recall all client caps (?) because dirty/pinned metadata is not flushed
Testing tasks.cephfs.test_client_limits.TestClientLimits.test_client_pin with this patch:... Patrick Donnelly
10:06 AM Backport #21602 (Fix Under Review): luminous: ceph_volume_client: add get, put, and delete object...
https://github.com/ceph/ceph/pull/18037 Ramana Raja
09:12 AM Backport #21602 (Resolved): luminous: ceph_volume_client: add get, put, and delete object interfaces
Wrap low-level rados APIs to allow ceph_volume_client to get, put, and
delete objects. The interfaces would allow Op...
Ramana Raja
09:07 AM Feature #21601: ceph_volume_client: add get, put, and delete object interfaces
https://github.com/ceph/ceph/pull/17697 Ramana Raja
09:06 AM Feature #21601 (Resolved): ceph_volume_client: add get, put, and delete object interfaces
Wrap low-level rados APIs to allow ceph_volume_client to get, put, and
delete objects. The interfaces would allow Op...
Ramana Raja
12:14 AM Backport #21600 (In Progress): luminous: mds: client caps can go below hard-coded default (100)
Patrick Donnelly
12:10 AM Backport #21600 (Resolved): luminous: mds: client caps can go below hard-coded default (100)
https://github.com/ceph/ceph/pull/18030 Patrick Donnelly
12:03 AM Bug #21575 (Pending Backport): mds: client caps can go below hard-coded default (100)
https://github.com/ceph/ceph/pull/16036/commits/538834171fe4524b4bb7cffdcb08c5b13fe7689f Patrick Donnelly

09/28/2017

10:29 AM Bug #4829 (Closed): client: handling part of MClientForward incorrectly?
Zheng Yan
10:20 AM Bug #6473 (Can't reproduce): multimds + ceph-fuse: fsstress gets ENOTEMPTY on final rm -r
open new one if we encounter this again Zheng Yan
10:17 AM Bug #12777 (Resolved): qa: leftover files in cephtest directory
Zheng Yan
10:16 AM Bug #11277 (Can't reproduce): hung fsstress run under thrash (no useful logs)
Zheng Yan
10:12 AM Bug #17212 (Resolved): Unable to remove symlink / fill_inode badness on ffff88025f049f88
Zheng Yan
10:07 AM Bug #20467 (Resolved): Ceph FS kernel client not consistency
Zheng Yan
10:06 AM Bug #21091 (Pending Backport): StrayManager::truncate is broken
https://github.com/ceph/ceph/pull/18019 Zheng Yan
10:04 AM Bug #19306 (Resolved): fs: mount NFS to cephfs, and then ls a directory containing a large number...
Zheng Yan
10:03 AM Bug #20313 (Resolved): Assertion in handle_dir_update
Zheng Yan
09:56 AM Bug #21168 (Resolved): cap import/export message ordering issue
Zheng Yan

09/27/2017

10:50 PM Bug #21584 (Fix Under Review): FAILED assert(get_version() < pv) in CDir::mark_dirty
Encountered this issue during snapshot test. The patch is from multimds snapsot fixes https://github.com/ceph/ceph/pu... Zheng Yan
10:40 PM Bug #21584 (Resolved): FAILED assert(get_version() < pv) in CDir::mark_dirty
... Zheng Yan
10:38 PM Bug #21551: Ceph FS not recovering space on Luminous
OK, it's likely caused by http://tracker.ceph.com/issues/19593. please don't enable standby reply for now Zheng Yan
01:55 PM Bug #21551: Ceph FS not recovering space on Luminous
This file system was create with Ceph v12.2.0. This cluster was cleanly installed with Ceph v12.2.0 and was never upg... Eric Eastman
03:36 AM Bug #21551: Ceph FS not recovering space on Luminous
... Zheng Yan
05:35 PM Bug #21575 (Resolved): mds: client caps can go below hard-coded default (100)
This was caused by the MDS cache in mempool change:... Patrick Donnelly
02:44 PM Feature #118: kclient: clean pages when throwing out dirty metadata on session teardown
I think sage means clean dirty pages at the same time of cleaning dirty metadata. This haven't been implemented yet Zheng Yan
02:39 PM Bug #17370 (Can't reproduce): knfs ffsb hang on master
Zheng Yan
02:27 PM Bug #21337 (Fix Under Review): luminous: MDS is not getting past up:replay on Luminous cluster
https://github.com/ceph/ceph/pull/17994
patch is only for luminous. this bug has been fixed in master in another w...
Zheng Yan
01:39 PM Feature #21571 (Duplicate): mds: limit number of snapshots (global and subtree)
when there are hundreds of snapshots, metadata operations become slow. Besides, lots of snapshots can cause kernel pa... Zheng Yan
01:11 PM Bug #19593 (Fix Under Review): purge queue and standby replay mds
https://github.com/ceph/ceph/pull/17990 Zheng Yan
09:45 AM Bug #19593: purge queue and standby replay mds
Ouch, It seems this one hasn't been fixed. this one can explan #21551
Zheng Yan
11:28 AM Bug #21568 (Resolved): MDSMonitor commands crashing on cluster upgraded from Hammer (nonexistent ...
Opened from mailing list thread "[ceph-users] "ceph fs" commands hang forever and kill monitors"... John Spray
09:32 AM Bug #20129 (Resolved): Client syncfs is slow (waits for next MDS tick)
Zheng Yan

09/26/2017

01:43 PM Bug #21551: Ceph FS not recovering space on Luminous
I uploaded the new mds run with
debug_mds=5
debug_journaler=10
to:
ftp://ftp.keepertech.com/outgoing/eric...
Eric Eastman
08:19 AM Bug #21551: Ceph FS not recovering space on Luminous
there are lots of "mds.0.purge_queue _consume: not readable right now" in the log.looks like purge queue stayed in n... Zheng Yan
04:07 AM Bug #21551: Ceph FS not recovering space on Luminous
The command 'ceph daemon mds.ede-c1-mon01 dump cache /tmp/cachedump' did not give any output so I ran
ceph daemon md...
Eric Eastman
02:08 AM Bug #21551: Ceph FS not recovering space on Luminous
Could you please run 'ceph daemon mds.ede-c1-mon01 dump cache /tmp/cachedump' and upload cachedump. Besides, please s... Zheng Yan
01:35 PM Bug #21433 (Closed): mds: failed to decode message of type 43 v7: buffer::end_of_buffer
great Zheng Yan
 

Also available in: Atom