Activity
From 10/16/2011 to 11/14/2011
11/14/2011
- 10:39 PM Revision 0276eab4 (ceph): rgw: adjust error code in swift copy failures
- 09:55 PM Revision 1fe16923 (ceph): rgw: fix swift responses encoding
- 09:23 PM Revision 2445fd84 (ceph): rgw: Fix some merge problems uncovered by gcc warnings:
- * a refactor in e2100bce left the mod_ptr and unmod_ptr members set
incorrectly in RGWCopyObj::init_common
* a fi... - 09:23 PM Revision cd900612 (ceph): Resolve gcc warnings.
- These should have no functional changes:
* Check errors from functions that currently cannot return any
* Initializ... - 08:15 PM Revision a5b8c851 (ceph): osd: remove dead osd_max_opq code
- This is no longer used as of a while ago!
Signed-off-by: Sage Weil <sage@newdream.net> - 05:02 PM rgw Bug #1698 (Resolved): radosgw-admin log list returns invalid json when a log object was created w...
- Fixed, commit:2e195500b5d3a8ab8512bcf2a219a6b7ff922c97. Not logging entries with non-utf8 bucket name.
- 04:30 PM Bug #1676: stats mismatch during snaps workunit
- Still happening in 11/11 nightly
1812/remote/ubuntu@sepia69.ceph.dreamhost.com/log/osd.1.log.gz - 04:27 PM Bug #1530: osd crash during build_inc_scrub_map
- Still happening in 11/11 nightly
1814/remote/ubuntu@sepia55.ceph.dreamhost.com/log/osd.1.log.gz - 04:24 PM Bug #1722: osd_class_dir must reflect autoconf libdir
- the original commit is commit:7e5dee907a8218647a88d1c7d3316cc277e1c44b. iirc that approach didn't work because autom...
- 02:11 PM Bug #1722: osd_class_dir must reflect autoconf libdir
- See also #1614, which for some reason doesn't let me edit it anymore.
- 02:11 PM Bug #1722 (Resolved): osd_class_dir must reflect autoconf libdir
- These two end up at different values for systems using /usr/lib64:
src/common/config_opts.h:285:OPTION(osd_class_d... - 04:19 PM Bug #1614 (Duplicate): default rados class location needs to be depend on autoconf libdir
- 02:08 PM Bug #1614: default rados class location needs to be depend on autoconf libdir
- 04:18 PM Revision f418775d (ceph): workunits: rados python workunit should be executable
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 04:12 PM Bug #1659: Upgrade from 0.27 -> 0.37 going wrong, OSDs miss map updates
- I saw a very similar stack trace in the 11/11 Nightly
1862/remote/ubuntu@sepia9.ceph.dreamhost.com/log/osd.5.log.gz
- 04:06 PM Revision b43981b8 (ceph): multimon: need at least 2 osds to go healthy
- 04:04 PM Bug #1724: timeout during tiobench test
- (I should have said the other problem was filed as bug 1723)
- 04:03 PM Bug #1724 (Resolved): timeout during tiobench test
- During the 11/11 nightlies, the tiotest task blocked multiple times. The first stack trace
(from 1831/remote/ubuntu... - 03:57 PM Bug #1723 (Rejected): timeouts during ffsb
- During the 11/11 nightlies, in suite 1827, sepia65 experienced multiple timeout events.
The first (from 1827/remote/... - 12:34 PM rgw Bug #1721 (Can't reproduce): rgw: spurious multipart-upload failures
- 11:56 AM Bug #1707 (Resolved): After fresh install, OSD initialization fails with: error error 17: File ex...
- great, thanks!
- 03:53 AM Bug #1707: After fresh install, OSD initialization fails with: error error 17: File exists not ha...
- Yes, I tested after that revision and could not reproduce the problem.
11/13/2011
- 10:18 PM Revision 102c4342 (ceph): crush: send debug output to dout, not stdout/err
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:16 PM Revision 25eee416 (ceph): test/run_cmd: use mkstemp instead of mkstemps
- my box didn't have mkstemps
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 10:07 PM Revision 18009866 (ceph): ceph-authtool: fix clitests
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 02:20 PM Bug #1688: Benjamin: pg stuck in scrub
- is this addressed by the pg lock vs transaction submit ordering changes?
- 02:13 PM Bug #1707: After fresh install, OSD initialization fails with: error error 17: File exists not ha...
- I think this is fixed by commit:7fb182a17b703002c1bd098391fb688b5b1e2749. Can you retest against latest master?
- 02:06 PM Bug #1708 (Can't reproduce): mon/PGMonitor.cc: 218: FAILED assert(paxos->get_version() + 1 == pen...
- I fixed a number of bugs in this area, and there was a big refactor. Can you retest the latest and see if you run in...
- 02:05 PM Feature #1720 (Duplicate): qa: rpm autobuilders
- probably start with opensuse and fedora, but eventually we probably want
- fedora (+ rawhide)
- opensuse (+ tumbl...
11/12/2011
- 11:17 PM Revision d476ae25 (ceph): test_str_list: make sure ' ' and ', ' separaters work for str lists
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:55 PM Revision ecd713c5 (ceph): ceph-authtool: make error msg more helpful
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:55 PM Revision 4f39aaa7 (ceph): keyring: don't print auid if it is the default
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:55 PM Revision ee02a1e1 (ceph): mon: implement 'fsid' command
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:19 PM Revision 5a3004e2 (ceph): Merge branch 'stable'
- 10:08 PM Revision 73f99a18 (ceph): mon: fix 'osd crush add ..' weight
- This was changed to floating point in commit 3f67893.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 10:05 PM Revision 1b843e0e (ceph): osdmap: build_simple with normal osd/host/rack/pool hierarchy
- This will be useful in the general case where the cluster is created with
an empty map and useful crush hierarchy.
S... - 10:04 PM Revision ec97c852 (ceph): mon: fix 'osd crush add ..' weight
- This was changed to floating point in commit 3f67893.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 09:42 PM Revision 0349fa96 (ceph): vstart.sh: don't generate initial osdmap explicitly
- This is simpler and exercises the monitors ability to start with a generic
osdmap and build it out as new osds are ad... - 09:41 PM Revision 30ddc85e (ceph): mon: make initial osdmap optional
- If an initial osdmap is not provided, we generate an empty one. The user
add osds on their own after that.
Signed-o... - 09:41 PM Revision 0d812252 (ceph): osdmap: build_simple: create reasonable pools when numosd==0
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:16 PM Revision 8e150fb4 (ceph): mon: add '--fsid foo' arg for setting generated monmap fsid
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:04 AM Revision b51d817e (ceph): mon: take '--fsid foo' arg with --mkfs
- This will set the seed monmap's fsid. This is useful if the monmap is
dynamically generated (e.g., based on ceph.con... - 05:04 AM Revision 0c731ed7 (ceph): osd: fix warnings
- osd/ReplicatedPG.cc: In member function 'virtual void ReplicatedPG::remove_watchers_and_notifies()':
osd/ReplicatedPG... - 04:52 AM Revision 73705f66 (ceph): monmaptool: fix clitests
- Initial map is epoch 0. Modifications still bump epoch by one.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:49 AM Revision 36241da4 (ceph): paxos: discard waiting_for_active events on reset
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:48 AM Revision 2253c016 (ceph): use libuuid for fsid
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:48 AM Revision 80ab6568 (ceph): monclient: use blank fsid (instead of epoch==0) for monmap checks
- We can safely mkfs with an epoch=0 monmap as long as the fsid is set. And
that is what commit f31825cee5300c708800a0...
11/11/2011
- 10:59 PM Revision 07950bb8 (ceph): crush: grammer: allow '.' in name token
- These are now in the generated crush maps, so it seems appropriate to
recompile them :).
Reported-by: Martin Mailand... - 10:54 PM Revision cf0a53e1 (ceph): mon: fix seed monmap removal
- Remove if we previous had no latest, not based on which map we now have.
It's possible we join when monmap epoch is s... - 10:52 PM Revision 6d370f3b (ceph): mon: allow monitor to automagically join cluster
- If a monitor starts up with the correct fsid and auth keys, it will now
add itself to the monmap (and subsequently tr... - 08:52 PM Revision d56485a8 (ceph): osd: pass monclient::init errors up the stack
- Fixes crash like
ceph version 0.38-149-gbf254de (commit:bf254de5cf8a17ce9467d166d87f3ab93170ae13)
1: (ceph::BackTr... - 08:37 PM Revision bf254de5 (ceph): mon: verify fsid during probe and election
- This will keep mismatched fsids out of the same quorum.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:22 PM Revision f1a98fb8 (ceph): mon: tolerate won election while active
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:22 PM Revision cd736b9d (ceph): mon: clean up logic a bit
- More explicit.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:22 PM Revision 2633d71d (ceph): mon: only re bootstrap if monmap actually changes
- If we go thru here just to update latest, that's fine; no need to restart
the bootstrap process.
Signed-off-by: Sage... - 08:15 PM Revision 622fbadd (ceph): paxos: fix off-by-one in share_state
- We hit this on adding a new monitor to an existing cluster.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:05 PM Revision 6c663d85 (ceph): mon: fix monmap update
- It's on the stack; update in place.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:02 PM Revision 1134fdfe (ceph): mon: properly process monmaps even when i have the latest
- We may get the latest monmap when we are doing our probing, but we still
need to process it in update_from_paxos(). ... - 07:55 PM Revision c097e634 (ceph): mon: fix up update_from_paxos() methods
- Make sure they behave when the initial state is learned from paxos.
Signed-off-by: Sage Weil <sage@newdream.net> - 07:41 PM Revision aea7563f (ceph): mon: create initial states after quorum is formed
- Signed-off-by: Sage Weil <sage@newdream.net>
- 07:41 PM Revision e545af2d (ceph): mon: remove empty monstore dirs
- This is sloppy, but it works well enough since we mkdir dirs as needed
too.
Signed-off-by: Sage Weil <sage@newdream.... - 07:41 PM Revision 65f797ea (ceph): mon: clean up mkfs seed data
- And make sure the monmap/latest gets written properly.
Signed-off-by: Sage Weil <sage@newdream.net> - 07:41 PM Revision f31825ce (ceph): monmaptool: new maps get epoch 0
- Just for consistency's sake.
Signed-off-by: Sage Weil <sage@newdream.net> - 06:45 PM Revision 1533f1c0 (ceph): mon: stage mkfs seed info in mkfs/ dir
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:34 PM Revision 9e941c43 (ceph): mon: eliminate PaxosService::init()
- update_from_paxos() is sufficient
Signed-off-by: Sage Weil <sage@newdream.net> - 06:19 PM Revision 0a926ef5 (ceph): mon: include monmap dump in mon_status and quorum_status
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:15 PM Revision 8c3d872e (ceph): mon: pull initial monmap from monmap/latest OR mkfs/monmap
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:05 PM Revision 0ecae996 (ceph): mon: take explicit initial monmap -or- generate one via MonClient
- This will simplify bootstrapping a cluster via e.g. mon_host.
Signed-off-by: Sage Weil <sage@newdream.net> - 09:58 AM Linux kernel client Bug #1704 (Resolved): oid limited to 40 chars, rbd images can be longer
- fixed by commit:224736d9113ab4a7cf3f05c05377492bd99b4b02
still need to do some cleanup here - 09:57 AM Linux kernel client Bug #1696 (Resolved): kclient: crash in ceph_d_prune
- fixed by commit:774ac21da76f5c3018428725074e27a3fd40b128
- 07:17 AM rgw Bug #1719 (Resolved): rgw: crash in ObjectCache::touch_lru
- ...
- 05:36 AM Revision 2bad0115 (ceph): filestore-idempotent
- run filestore_idempotent.py task.
- 05:35 AM Revision c5f070b8 (ceph): filestore_idempotent.py: simple task to test non-idempotent osd ops
- Write some non-idempotent events to the osd. Simulate a failure. Verify
the result is correct on replay.
This must... - 05:12 AM Revision 69cd3625 (ceph): filestore: sync after non-idempotent operations
- This is a big hammer to fix journal replay on non-btrfs fs backends (extN,
xfs, whatever). The problem is that it is... - 05:12 AM Revision 09811120 (ceph): filestore: document the btrfs_* fields
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:12 AM Revision 8df0cd38 (ceph): filestore: make trigger_commit() wake up sync; adjust locking
- We need to wake up the sync thread (duh).
Also, we need to obey the FileJournal::lock -> journal_lock locking
order.... - 05:12 AM Revision 9f1673c1 (ceph): test_filestore_idempotent: transactions are individually idempotent
- Make individual transactions idempotent, but their interactions
non-idempotent. I.e. A A A A is okay, but A B A is n... - 05:12 AM Revision add04d15 (ceph): filejournal: fix replay of non-idempotent ops
- - start sync thread prior to replay, so that we can commit as we replay
operations
- keep applied_seq accurate
- pa... - 05:12 AM Revision dae6c956 (ceph): test_filestore_idempotent: detect commit cycles due to non-idempotent ops
- If we do a non-idempotent op and it does a commit itself, we don't see
fs->is_committed() true ever. Also count full... - 04:50 AM Revision fa5047b3 (ceph): Merge remote branch 'gh/stable'
- 01:15 AM Revision 1c1ebb4d (ceph): Add rados python tests.
- 01:10 AM Revision 2fb70297 (ceph): rgw: remove warning
- 01:03 AM Revision 5407fa70 (ceph): workunits: add workunit for running rgw and rados python tests
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:52 AM Revision 71bfe897 (ceph): test/pybind: add test_rgw
- Forgot to add this in the previous commit.
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> - 12:46 AM Revision ea42e02c (ceph): test/pybind: convert python rados and rgw tests to be runnable by nose
- These tests can now be run automatically more easily.
Fixes: #1653
Signed-off-by: Josh Durgin <josh.durgin@dreamhost... - 12:37 AM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Yes I am stopping the clients and remounting...but if im doing a mkcephfs, i make sure to umount all the clients befo...
- 12:33 AM Revision 25cde7f9 (ceph): rados.py: fix Snap.get_timestamp
- This now uses datetime, imports the right things, and calls the right function.
Fixes #1577
Signed-off-by: Josh Durg...
11/10/2011
- 11:07 PM Revision b600ec2a (ceph): v0.38
- 11:05 PM Revision 2a7fbe0c (ceph): common: return null if mc.init() unsuccessful
- Prevents ceph.cc from segfaulting on missing keyring.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 11:05 PM Revision a177a702 (ceph): rbd.py: fix list when there are no images
- It should return [], not [''].
Reported-by: Eric Chen <Eric_YH_Chen@wistron.com>
Signed-off-by: Josh Durgin <josh.du... - 11:05 PM Revision 27bb48c5 (ceph): mon: overwrite in put_bl
- This fixes a situation where we accept a large value, there is some failure
and recovery, and then we commit a smalle... - 11:05 PM Revision 2f97a222 (ceph): PG: mark scrubmap entry as not absent when we see an update
- Previously, there would be an assert failure in _scan_list if we see an
object deleted and then recreated.
Signed-of... - 10:58 PM Revision 87941128 (ceph): rgw: implement swift copy, fix copy auth
- 10:13 PM Revision 77c977c1 (ceph): misc: allow >1 monitor per role in get_mon_names()
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:09 PM Revision 704644bc (ceph): PG: gen_prefix: use osdmap_ref rather than osd->osdmap
- Otherwise, the debug output might not match the map used by
the pg logic.
Signed-off-by: Samuel Just <samuel.just@dr... - 10:09 PM Revision 7fb182a1 (ceph): OSD: sync_and_flush afer mkfs to create first snap
- Previously, if we kill the OSD process before the filestore
does its first sync, we end up replaying the journal on t... - 09:41 PM Bug #1670 (Can't reproduce): osd: crash in update_heartbeat_peers
- 09:38 PM Bug #213 (Resolved): non-idempotent transactions (clone) under ext3 may not replay correct result
- commit:dae6c956543276e103a272eb1e897db17b840348
- 08:54 PM Bug #1530: osd crash during build_inc_scrub_map
- 05:29 PM Bug #1530: osd crash during build_inc_scrub_map
- We just found surprisingly similar stack traces in three of last night's failures:
nightly_coverage_2011-11-10/1740/... - 06:45 PM Feature #1516 (Resolved): openstack: single node dev environment
- 05:06 PM rgw Feature #1717 (Resolved): rgw: support json input
- 05:06 PM Feature #1653 (Resolved): librados: python binding nose tests
- Fixed by commit:ea42e02ca2fd3655dbaf2e720e31d78da5022e21.
- 05:05 PM rgw Cleanup #1716 (Closed): rgw: remove curl use
- 05:05 PM Bug #1577 (Resolved): rados.py: Snap.get_timestamp does not work
- Fixed by commit:25cde7f98ac195b0458830a3e345db54a994384b.
- 04:57 PM Feature #1539 (Duplicate): libvirt: make sure snapshots work
- 04:11 PM rgw Feature #1715 (Rejected): rgw: use RENAME osd operation to avoid slow CLONE operations
- add to osd too
- 04:03 PM rbd Feature #1713 (Resolved): teuthology: qemu tasks, tests
- gitbuilder
teuthology task
some tests that run in it - 03:29 PM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Gokul Krishnan wrote:
> Thank you for reverting back so quickly.
>
> Well in my scenario, i just have one Ceph se... - 03:29 PM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Gokul Krishnan wrote:
> by the way,
> you have assigned a target version as v0.39...but in the site i can find only... - 01:50 AM CephFS Bug #1702: Ceph MDS crash + client mount problem
- by the way,
you have assigned a target version as v0.39...but in the site i can find only the source for v0.37...
e... - 12:45 AM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Thank you for reverting back so quickly.
Well in my scenario, i just have one Ceph server running. And yes, every ... - 03:29 PM rgw Feature #1712 (Resolved): rgw: support swift manifest objects
- 03:22 PM Feature #1711 (Resolved): chef: multiple monitor support
- 03:22 PM Bug #1669 (Resolved): linux 32 bit kernel client ld libraries and rm issue
- 03:14 PM Feature #1709 (Resolved): specfile: merge suse spec file changes
- 03:00 PM rgw Bug #1706 (Resolved): rgw: copy object auth verification (probably) broken
- 02:59 PM rgw Bug #1706: rgw: copy object auth verification (probably) broken
- Fixed, commit:87941128b60608d66dc5327038f099a1fb2a99c3.
- 02:59 PM rgw Bug #1705 (Resolved): rgw: swift copy is broken
- Fixed, commit:87941128b60608d66dc5327038f099a1fb2a99c3.
- 02:57 PM CephFS Feature #1448: test hadoop on sepia
- The following benchmark, TestDFSIO, is for 12 OSDs, 1 MDS/MON. There is a single ext4 disk per node dedicated to Ceph...
- 02:46 PM Bug #1632 (Can't reproduce): osd: crash in dequeue_op
- 01:54 PM Bug #1708 (In Progress): mon/PGMonitor.cc: 218: FAILED assert(paxos->get_version() + 1 == pending...
- 01:45 PM Bug #1708 (Resolved): mon/PGMonitor.cc: 218: FAILED assert(paxos->get_version() + 1 == pending_in...
- Running ceph version from git: a3dd5bd67ba19aae51a51318138ef10213a91449
Slaves are all ubuntu 11.10, 3.0.0-12
Files... - 12:06 PM Bug #1707 (Resolved): After fresh install, OSD initialization fails with: error error 17: File ex...
- Running ceph from git @ a3dd5bd6 with btrfs
Ubuntu 11.10, 3.0.0-12 on all machines
After installing my compiled c... - 01:17 AM Revision a3dd5bd6 (ceph): PG: update info.history even if lastmap is absent
- Previously, we did not update same_interval_since etc if
we do not have the previous map.
Signed-off-by: Samuel Just... - 12:36 AM Revision 023ff590 (ceph): Makefile: add MMonProbe.h
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:33 AM Revision fd5fb993 (ceph): osd: remove useless proc_replica_log() side-effect
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
11/09/2011
- 11:38 PM Revision 78ad144a (ceph): hadoop: update patch and Readme.
- Patch generated by Noah Watkins <noahwatkins@gmail.com>
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 11:30 PM Revision 386c0db3 (ceph): rgw: swift guesses mime type if not specified
- 10:50 PM Revision 78ccb2a9 (ceph): osd: comment PG::lock*(), whitespace
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:46 PM Revision 87318389 (ceph): Merge branch 'master' of github.com:NewDreamNetwork/ceph
- Conflicts:
src/osd/PG.cc - 10:32 PM Revision 5fa8df1e (ceph): osd: improve last_peering_reset debugging
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:32 PM Revision 383dfa33 (ceph): crypto: make crypto handlers non-static
- These were static in auth/Crypto.cc, which was mostly fine, except when
we got a signal shutting everything down for ... - 10:15 PM Revision 9db994a5 (ceph): PG: always add backlog entry
- Previously, we did not add a backlog entry if the object already had an
entry in the log along with an entry for that... - 10:15 PM Revision 0dffddf3 (ceph): osd/: change type of osd::osdmap to a shared_ptr
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 10:15 PM Revision 5df28ece (ceph): OSDMap,CrushWrapper: const cleanup on OSDMap
- The osd's cached maps are not actually modified once cached. Marking
these methods const (which they should be) allo... - 10:15 PM Revision b41b1fa5 (ceph): PG: cache read-only reference to the current osdmap on pg lock
- Previously, we needed to grab an osd_map read lock to send messages,
among other things. Now, we grab a reference to... - 10:04 PM Revision 15da4787 (ceph): rbd: Fix the showmapped cmd usage
- If the rbd showmapped cmd is given any extra arguments, rbd will fail
with "assert(0)". Fix it by exiting with "usage... - 09:37 PM Revision 303e863d (ceph): add hammer.sh
- simple script to repeat a test until it fails. can probably do something much more sophisticated
here, but this works. - 09:28 PM Revision 33549333 (ceph): hadoop: return all replica hostnames
- Updates CephFileSystem to return all replica locations,
and in addition attempts to use reverse DNS to convert
the OS... - 09:23 PM Revision e6035a62 (ceph): hadoop: make listStatus quiet
- Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
- 09:23 PM Revision d7f911fb (ceph): hadoop: handle new ceph_get_file_stripe_address
- Updates the Hadoop JNI/CephFileSystem to handle
the new version of ceph_get_file_stripe_address
which returns the loc... - 09:23 PM Revision 619430a7 (ceph): client: return stripe address replicas
- Changes ceph_get_file_stripe_address to return a
vector of entity_addr_t's for the primary and the
replicas. libcephf... - 09:15 PM Revision c5c50377 (ceph): client: fix bad perfcounter fset callers
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:50 PM Revision 808c6442 (ceph): Improve use of syncfs.
- Test syncfs return value and fallback to btrfs sync and then sync.
Signed-off-by: Alexandre Oliva <oliva@lsd.ic.unic... - 08:48 PM Revision c51e2f72 (ceph): osd: fix perfcounter typo
- Signed-off-by: Sage Weil <sage@newdream.net>
- 07:43 PM Revision 1ac6b47c (ceph): os: rename and make use of the split_threshold parameter.
- This was accidentally left out of the must_split calculation. Put it
in, and rename it to split_multiplier (as that i... - 07:03 PM Revision 09455eea (ceph): perfcounters: fix users of fset on averages
- I forgot to audit these before merging the assert and they popped up
in teuthology and stuff. :(
Signed-off-by: Greg... - 06:49 PM Revision afa56f16 (ceph): nuke: increase reboot timeout
- Some sepia nodes are very slow to reboot.
- 05:35 PM Bug #1690: osd re-created from scratch will crash on start-up
- I was using v0.37; in order to debug this, I first build top of the tree stable (b8979f4d292f6a739daac81ce8e59aa084e1...
- 05:11 PM rgw Bug #1706 (Resolved): rgw: copy object auth verification (probably) broken
- Looking at RGWCopyObj::verify_permission(), we don't look at the source acl, but rather at the source bucket's acl.
- 05:07 PM rgw Bug #1705 (Resolved): rgw: swift copy is broken
- Swift can accept alternative HTTP COPY method (with src/dest transposed).
- 04:38 PM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- 02:55 PM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- Update: the current first pass plan is to initiate a FileStore sync after any non-idempotent operation. This updates...
- 03:35 PM Linux kernel client Bug #1701: krbd: limits and constants are not consistent in kernel and userspace
- Also related: we have MAX_POOL_NAME_SIZE and MAX_SNAP_NAME_SIZE as 128 in qemu right now.
- 02:37 PM Linux kernel client Bug #1701: krbd: limits and constants are not consistent in kernel and userspace
- Stratos Psomadakis wrote:
> Instead of opening a new issue, I think I can add it here.
>
> Besides those limits o... - 02:18 PM Linux kernel client Bug #1701: krbd: limits and constants are not consistent in kernel and userspace
- Instead of opening a new issue, I think I can add it here.
Besides those limits on the RBD images, there's also a ... - 12:44 PM Linux kernel client Bug #1701 (New): krbd: limits and constants are not consistent in kernel and userspace
- There are a few things that exist in the kernel but not userspace:
* SNAP_NAME_LEN
* (MIN|MAX)_OBJECT_ORDER
Also... - 03:00 PM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Ok, so generally speaking, the only time you shoudl see fsid mismatches like that is if you have daemons from multipl...
- 02:55 PM CephFS Bug #1702: Ceph MDS crash + client mount problem
- Hello,
thank you for the reply.
no, unfortunately i am not able to reproduce the error using debug ms = 20(for MD... - 01:23 PM CephFS Bug #1702 (Need More Info): Ceph MDS crash + client mount problem
- Are you able to reproduce this with 'debug mds = 20' and 'debug ms = 20' in your ceph.conf [mds section]?
Not sure... - 12:51 PM CephFS Bug #1702 (Can't reproduce): Ceph MDS crash + client mount problem
- Hello,
i have configured ceph using a configuration as shown here[[http://pastebin.com/sQb8WZbx]].
The Ceph serve... - 02:43 PM Bug #1684 (Duplicate): mon: crash in CryptoKey::encrypt
- 02:42 PM Bug #1633 (Resolved): osd crash in CryptoKey::decrypt
- should be fixed by commit:383dfa33682abeae7348655fc103dd80c41b7ba7
- 02:39 PM Linux kernel client Feature #962 (Resolved): d_prune
- 02:39 PM Linux kernel client Bug #850 (Resolved): make NULL lookup using I_COMPLETE work
- 02:39 PM Linux kernel client Bug #851 (Resolved): make dcache readdir with I_COMPLETE work
- 02:38 PM Linux kernel client Bug #1704 (Resolved): oid limited to 40 chars, rbd images can be longer
- From Stratos Psomadakis:
"Besides those limits on the RBD images, there's also a hardcoded limit in
libceph (mess... - 02:27 PM rgw Bug #1698: radosgw-admin log list returns invalid json when a log object was created with a name ...
- This is my vote for "let's not allow radosgw clients to create artifacts with non-utf8 names in the first place".
- 02:19 PM Bug #1530 (Resolved): osd crash during build_inc_scrub_map
- 02:08 PM Bug #1703 (Resolved): rbd: showmapped cmd fails, when extra args are present
- 02:00 PM Bug #1703 (Resolved): rbd: showmapped cmd fails, when extra args are present
- rbd showmapped cmd will fail with assert(0), when given any extra arguments.
Patch to fix it attached (exiting wit... - 01:02 PM Bug #1695 (Rejected): wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- Serge Rittscher wrote:
> ok, the output is:
> @
> rm -f init-ceph init-ceph.tmp
> sed -e 's|@bindir[@]|/usr/local... - 11:39 AM Bug #1695: wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- ok, the output is:
@
rm -f init-ceph init-ceph.tmp
sed -e 's|@bindir[@]|/usr/local/bin|g' -e 's|@libdir[@]|/usr/lo... - 11:04 AM Bug #1695: wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- oops, 'touch init-ceph.in' first, then 'make init-ceph'
- 12:49 AM Bug #1695: wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- @make init-ceph@
returns:
@make: `init-ceph' is up to date.@
- 11:10 AM Bug #1700 (Resolved): osd: invalid perfcounter usage
- Should be fixed in commit:09455eeac4fb37c31998202ad9503901f53c21dc. My bad!
- 10:14 AM Bug #1700 (Resolved): osd: invalid perfcounter usage
- During dbench, two osds crashed on this assert:...
- 11:09 AM Bug #1694 (Resolved): monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- 11:09 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- oh nevermind, didn't see that second comment. the fix is commit:0bcdd4f3b2a2dba405639122b84f7aad978f347b, which come...
- 11:06 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- Great. Can you attach (or email) the ceph.conf you're using?
Thanks! - 07:55 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- The monitor that was generating the osdmap was running commit:5bd029ef01fcb59bea9170af563c3499cce1e8c4 and that faile...
- 02:25 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- Ok, I've ran those commands and it gives me:...
- 07:19 AM CephFS Bug #1472: cfuse hangs with v0.34
- Some of the hangs we've been seeing on the client may have been related to having two nics on each node. We had seen...
- 06:17 AM Revision 6d39cc11 (ceph): ceph: keep ceph.conf at ctx.ceph.conf
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:17 AM Revision 60863f70 (ceph): ceph_manager: manipulate monitors
- 06:17 AM Revision 6618a027 (ceph): mon_recovery: add task to test monitor cluster failure recovery
- Some simple tests to start with. We still need some sort of mon cluster
thrashing.
Signed-off-by: Sage Weil <sage@n... - 06:16 AM Revision 9acea7a6 (ceph): multimon mon_recovery tests on variously sized monitor clusters
- 06:11 AM Revision 6ab14874 (ceph): Merge branch 'wip-mon'
- 05:58 AM Revision 87634ce1 (ceph): osd: don't open deleted map from generate_past_intervals
- The first get_map() call needs to be avoided when stop < last_epoch. This
fixes a crash like
2011-11-08 21:51:09.04... - 05:13 AM Revision 20cf1e96 (ceph): automake: enable 'make V=0'
- Enables silent mode for automake generated Makefiles,
and silent mode is _off_ by default. Using V=0 the output
is mu... - 12:45 AM Revision 4b0cf89b (ceph): Add rbd python binding test.
- 12:24 AM Revision 1bc1a244 (ceph): mon: handle active -> electing transition properly
- If we are already active, make sure we reset things properly before going
into an election.
Signed-off-by: Sage Weil... - 12:09 AM Revision 5d32bcae (ceph): Add nuke-on-error option.
- This lets automated jobs nuke and unlock machines after failed
tests. Each machine is nuke individually, so one down ... - 12:09 AM Revision 006a0dd4 (ceph): Remove unused imports and variable.
11/08/2011
- 10:21 PM Feature #1007 (Resolved): qa: osd failure and cluster recovery test(s)
- yay thrashing
- 10:20 PM Bug #1694 (Need More Info): monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- 09:28 PM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- Can you try this and see if there is a mismatch?...
- 10:06 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- Aha! Read that wrong, tnx.
I used mkcephfs to generate the crushmap, I did not write my own. - 09:17 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- max_osd in the osdmap needs to be >= the max_devices in the crush map. how did you set up the cluster? did mkcephfs...
- 07:18 AM Bug #1694: monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- I just made a small adjustment to crushtool so it would print max_devices:...
- 07:01 AM Bug #1694 (Resolved): monitor crash: FAILED assert(get_max_osd() >= crush.get_max_devices())
- I just did a fresh install of my cluster and after starting I saw my monitors go down with:...
- 10:18 PM Feature #1646 (Resolved): mon: catch up on committed items before attempting to join quorum
- 10:17 PM Revision 7a32cc60 (ceph): rgw: swift bucket report returns both bytes size and actual size
- 10:17 PM Revision 76090324 (ceph): rgw: don't return partial content response with bad header
- 10:17 PM Revision a04afd09 (ceph): rgw: abort early on incorrect method
- 09:33 PM Bug #1695: wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- What is the output if you...
- 09:06 AM Bug #1695 (Rejected): wrong path to ceph's libs / bash scripts in /etc/init.d/ceph
- After installing Ceph from sources (version ceph-0.37.tar.gz) on Ubuntu by executing
$ ./autogen.sh
$ ./configure... - 09:09 PM Revision 2fb73bdd (ceph): paxos: fix race between active and commit
- If paxos reproposes an old learned value, we have a C_Active waiter, and
also a commit in progress.
When we reach qu... - 08:56 PM Revision 1ffb7b97 (ceph): mon: add 'quorum_status' command
- Show status of the current quorum. Block until there is one.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:52 PM Revision a8b28ee5 (ceph): mon: do not participate in the election unless we are in electing state
- If we participate, we may be included in the quorum, even tho we are
probing, slurping, whatever.
Signed-off-by: Sag... - 07:50 PM Revision 64350c0b (ceph): rgw: guard perfcounter accesses in rgw_cache.
- This gets called by radosgw-admin, so it needs to handle
perfcounter being a null pointer.
Signed-off-by: Greg Farnu... - 07:28 PM Revision 42f5f024 (ceph): rgw: initialize all the perfcounters, in order
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 06:42 PM Revision e952e10f (ceph): ReplicatedPG: use finc, not fset, on average counters
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 06:42 PM Revision 29e091b5 (ceph): mon: 'mon_status' command to dump individual mon state
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:04 PM Revision f0b9a331 (ceph): rgw: use l_rgw_qactive perfcounter
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 05:58 PM Revision 9035ffb2 (ceph): mon: add probe+slurp timeouts
- A short timeout on probe, so we can form new quorums quickly.
A longer timeout on slurp, so we will tolerate a slow ... - 05:50 PM Revision 0fe0f9db (ceph): rgw: create and tear down a radosgw perfcounter
- Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 05:50 PM Revision d0b226e7 (ceph): perfcounter: assert when you try and set an average.
- If you're trying to set an average, you're probably doing it wrong.
Signed-off-by: Greg Farnum <gregory.farnum@dream... - 05:50 PM Revision 57b60b8a (ceph): perfcounter: add some minimal documentation.
- The data model is a bit obtuse if you're just looking at the code.
Signed-off-by: Greg Farnum <gregory.farnum@dreamh... - 05:50 PM Revision cf566550 (ceph): rgw: implement perfcounters
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 04:59 PM Linux kernel client Bug #1696: kclient: crash in ceph_d_prune
- Here is the code:...
- 11:50 AM Linux kernel client Bug #1696 (Resolved): kclient: crash in ceph_d_prune
- During the 11/08 nighly, several suites:
1606 autotest dbench
1607 workunit direct_io
1608 workunit kc... - 04:57 PM Bug #1684: mon: crash in CryptoKey::encrypt
- This happened on an mds during a thrashing run:...
- 04:29 PM Linux kernel client Feature #1699 (Resolved): debug symbols in autobuilt (sepia) kernels
- We need debug symbols in the .ko objects:...
- 03:49 PM rgw Bug #1698: radosgw-admin log list returns invalid json when a log object was created with a name ...
- The two preceding days show similar errors as well.
- 03:48 PM rgw Bug #1698: radosgw-admin log list returns invalid json when a log object was created with a name ...
- The description above is malformed for whatever reason, so I'll try again:
radosgw-admin log list is producing bad J... - 03:44 PM rgw Bug #1698 (Resolved): radosgw-admin log list returns invalid json when a log object was created w...
- 2011-11-07-12-0-<80>..
- 02:34 PM rgw Feature #1697 (Resolved): s3-tests: test bucket headers
- 12:04 PM rgw Feature #1591 (Resolved): rgw: instrument with perfcounter
- Finally sat down and did this. Merged in commit:64350c0b4d3ba2061cebed87f4cd6f513d2ba6ed and passed s3tests.
- 06:46 AM Revision 2523b70e (ceph): mon: slurp latest state from active monitors before joining quorum
- If a monitor has been down and is behind, and joins the quorum, the
other nodes will try to send it all of the needed... - 06:41 AM Revision c2fc986e (ceph): monmap: simplify constructor
- Explicitly set created, last_changed where appropriate.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 06:41 AM Revision 279661f3 (ceph): paxos: last_consumed == latest_stashed; behave accordingly
- Initialize on startup.
Don't re-read off of disk on every trim_to() call.
Signed-off-by: Sage Weil <sage.weil@dreamh... - 06:41 AM Revision 100fba8e (ceph): mon: fix osdmap trim
- We can raise the floor even when min_last_epoch_clean if very close to
the current version, as long as it is still ab... - 04:40 AM Revision 628de548 (ceph): mon: don't call out to mon->call_election for internal election restarts
- This lets us drop the is_new kludge.
Signed-off-by: Sage Weil <sage@newdream.net> - 04:40 AM Revision 18941dd0 (ceph): mon: rename election_starting -> restart
- These callbacks reset monitor/paxos/paxosesrvice state, which used to
happen when an election started, but will now n... - 04:40 AM Revision 2f46e8cd (ceph): mon: revamp monitor states
- starting -> probing, electing
some cleanup
Signed-off-by: Sage Weil <sage@newdream.net> - 04:40 AM Revision 40843eb3 (ceph): rgw: fix warning
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 01:08 AM Revision 2836104a (ceph): rgw: fix accept-range for suffix format, other related issues
11/07/2011
- 11:04 PM Revision 2f881e12 (ceph): Timer.cc: remove global thread variable
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 11:04 PM Revision d4ef9215 (ceph): common: return null if mc.init() unsuccessful
- Prevents ceph.cc from segfaulting on missing keyring.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 09:05 PM Revision c764b247 (ceph): Fix leftover orchestra import clause.
- This seems to be a leftover from
a2372fce12b6bd1818e155d1d8ed5134dbd8fd4a,
no idea how it stayed hidden this long. - 05:27 PM Revision 480b8260 (ceph): rbd: add showmapped to clitests and rst man page
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 05:27 PM Revision 4e518ed3 (ceph): rbd: Document the rbd showmapped cmd
- Document the rbd showmapped cmd in rbd.usage(), and rbd's man page,
and add it to the bash completion script.
Signed... - 05:10 PM Revision 34d80397 (ceph): rbd.py: fix list when there are no images
- It should return [], not [''].
Reported-by: Eric Chen <Eric_YH_Chen@wistron.com>
Signed-off-by: Josh Durgin <josh.du... - 03:35 PM Bug #1690: osd re-created from scratch will crash on start-up
- I seem to be having some trouble reproducing this. What version are you running? Could you repeat the procedure wit...
- 10:33 AM Bug #1690 (Can't reproduce): osd re-created from scratch will crash on start-up
- Some time ago, it was possible to re-create an osd after its filesystem failed as simply as running “cosd -i # --mkfs...
- 02:59 PM CephFS Feature #1693: libcephfs: Support TRIM (hole punching)
- Kernelside ceph.ko ticket is #591. Let this ticket stand for the userspace libcephfs (and ceph-fuse) support.
- 02:12 PM CephFS Feature #1693 (Resolved): libcephfs: Support TRIM (hole punching)
- 02:57 PM Feature #1692: librbd: Support TRIM (hole punching) (userspace client)
- Kernel-side rbd.ko ticket is #190. Let this ticket stand for the librbd (userspace) support.
- 02:11 PM Feature #1692 (Duplicate): librbd: Support TRIM (hole punching) (userspace client)
- 01:56 PM Bug #1691 (Can't reproduce): rados export failures
- ...
- 11:36 AM Linux kernel client Bug #1667 (Resolved): BUG at fs/inode.c line 1375
- 11:17 AM rbd Feature #1662 (In Progress): libvirt: obscure qemu/rbd secrets
11/06/2011
- 03:08 PM Linux kernel client Bug #1667: BUG at fs/inode.c line 1375
11/05/2011
- 09:37 PM Linux kernel client Bug #1686 (Resolved): directory not empty errors
- fixed commit:c6ffe10015f4e6fba8a915318b319c43aed1836f clear helper
- 09:37 PM Linux kernel client Bug #1687 (Resolved): directory existence failures
- fixed commit:c6ffe10015f4e6fba8a915318b319c43aed1836f clear helper
- 01:38 AM Revision ae41f323 (ceph): OSD: write_info/log before dropping lock in generate_backlog
- Bug #1530
This should fix the following race:
1) osd->generate_backlog does pg->assemble_backlog
2) osd->generate_ba... - 12:30 AM Revision fb70f5cc (ceph): FileJournal: stop using sync_file_range
- Using sync_file_range means that neither any required metadata gets commited,
nor the disk cache gets flushed. Stop ... - 12:29 AM Revision 585a46c5 (ceph): monclient: simplify auth_supported set
- Use AuthSupported class instead of repopulating it ourselves.
Signed-off-by: Sage Weil <sage@newdream.net> - 12:23 AM Revision a38c0054 (ceph): test_libcephfs
- 12:21 AM Revision 10141673 (ceph): Makefile: use static add for test_libcephfs_readdir.
- Otherwise it doesn't seem to play nicely with teuthology/sepia
due to requiring the host to have gtest installed.
Si...
11/04/2011
- 09:57 PM Revision 5b4e9d31 (ceph): RadosModel: add DeleteOp to test object deletions
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 08:40 PM Revision 280a4d1d (ceph): rgw: fix tmp objects leakage
- 08:13 PM Revision 8d914f0e (ceph): rgw: list system buckets through rados api
- 08:13 PM Revision fc6522a8 (ceph): rgw: don't purge pools in any case
- 06:44 PM Bug #1530: osd crash during build_inc_scrub_map
- ae41f3232a39dbf33487ab02cbac292f58debea8
- 04:59 PM Bug #1530: osd crash during build_inc_scrub_map
- My best guess about this bug goes something like this:
1) osd->generate_backlog does pg->assemble_backlog
2) osd->g... - 05:20 PM Linux kernel client Bug #1686: directory not empty errors
- this is probably due to the d_prune stuff i just pushed to master. need to do some serious debugging here.
the re... - 01:43 PM Linux kernel client Bug #1686 (Resolved): directory not empty errors
- Today, many of the kclient ceph fs tests failed due to problems removing directories. This did not happen with yester...
- 04:52 PM Bug #1689 (Can't reproduce): osd: segfault in recover_primary
- This happened in run 1497, thrashing with the snaps workload, on 3 osds....
- 04:50 PM Bug #1529: cosd: os/FileStore.cc: 2390: FAILED assert(0 == "ENOENT on clone suggests osd bug")
- Thrashing with the snaps workload triggered this on several osds in run 1497 today.
- 03:21 PM Bug #1683: librados: list objects should also return locator key
- Apparently, I implemented this about 2 months ago but didn't merge it...
- 01:19 PM Bug #1683 (Resolved): librados: list objects should also return locator key
- 02:47 PM CephFS Bug #1472: cfuse hangs with v0.34
- We're seeing similar hangs again. One thing I didn't mention in my previous posts, we are always adjusting the repli...
- 02:43 PM Bug #1688 (Closed): Benjamin: pg stuck in scrub
- Looks like the bug is related to last_update_applied not getting up to last_update on primary. No further scrubbing ...
- 01:46 PM Linux kernel client Bug #1687 (Resolved): directory existence failures
- Some benchmarks today failed to cd to directories. These worked yesterday.
From blogbench and ffsb:... - 01:40 PM rgw Bug #1685 (Resolved): rgw: tmp objects leakage
- Yes, but the problem was elsewhere. Fixed, commit:280a4d1ded4b83974805c60bcd410ee00ccc3884.
- 01:38 PM rgw Bug #1685: rgw: tmp objects leakage
- This is probably due to to #1683, as tmp objects are all placed using locators, right?
- 01:27 PM rgw Bug #1685 (Resolved): rgw: tmp objects leakage
- After running radosgw-admin temp remove, we're still left out with objects from the tmp namespace. Either we fail to ...
- 01:21 PM Bug #1684 (Duplicate): mon: crash in CryptoKey::encrypt
- From teuthology:~teuthworker/archive/nightly_coverage_2011-11-04/1472/teuthology.log:...
- 01:17 PM rgw Bug #1672 (Resolved): rgw: support chunked transfer encoding
- Done.
- 12:55 PM CephFS Bug #1682 (Resolved): mds: segfault in CInode::authority
- From teuthology:~teuthworker/archive/nightly_coverage_2011-11-04/1469/teuthology.log:...
- 12:30 PM rgw Bug #1681 (Resolved): rgw: user rm with --purge doesn't remove data
- I just disabled it as it did it incorrectly
- 10:25 AM Feature #1618: libvirt: make sure migration works
- Mike Lowe emailed me and mentioned it works for him on Oneiric with a custom kvm 0.15.1, no other changes. I still wa...
- 09:53 AM CephFS Feature #1680 (New): support reflink (cheap file copy/clone)
- It seems the API is still fs-specific ioctls, but there's repeated discussion about reflink(2).
If a nice common API... - 08:19 AM Bug #1679: assertion failure is_replica()
- Upon trying to restart the failed osds, other osds (7) fail:
*** Caught signal (Aborted) **
in thread 0x7fcceb... - 08:12 AM Bug #1679 (Can't reproduce): assertion failure is_replica()
- 3 boxes, 12 osds per box. 4 osds (9,11,20,24) crashed at the following assertion. This was triggered by first setti...
11/03/2011
- 11:01 PM Revision 0f98006c (ceph): rgw: fix PUT without content length (non chunked)
- 10:46 PM Revision 256ac72a (ceph): rbd: document --order and list required args where they're necessary
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 09:43 PM Revision 0df3f036 (ceph): Merge remote branch 'nwatkins/for-master'
- 09:11 PM Revision 90249069 (ceph): Merge branch 'wip-getdir'
- 08:59 PM Revision b8733476 (ceph): gitignore: just ignore all test_ files
- We don't want to add a new ignore for each test!
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 08:55 PM Revision d4faf588 (ceph): qa: workunit to run test_libcephfs_readder
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 08:49 PM Revision 120c3fbd (ceph): test: write a test to try and check on Client::readdir_r_cb.
- It's made difficult by having to go through libcephfs, but it's better
than nothing and should catch most of the erro... - 08:39 PM Feature #1678 (Resolved): rados tool: ability to specify object locator
- We need to be able to access objects with none-default locators.
- 08:27 PM Revision 4f3b1138 (ceph): ceph_manager: log ceph -s output so progress is visible in the logs
- 08:08 PM Revision 0b451f94 (ceph): Keep each ssh connection alive.
- With long-running jobs like thrashing, ssh connections were timing
out. - 08:07 PM Revision 6e3e0d7c (ceph): connection: allow the caller to specify whether keep-alive should be used
- 06:45 PM Revision 58eb8c5e (ceph): rgw: fix null deref, cleanups
- 06:29 PM Revision 0d4987d9 (ceph): rgw: fix crash when accessing swift auth without user
- 06:29 PM Revision 7726e78d (ceph): rgw: add support for chunked upload
- 06:29 PM Revision b1a0c1ad (ceph): locker: fix race in locking
- The isolation level is lower than I thought. This made it possible for
two clients to think they both locked the same... - 04:39 PM CephFS Bug #1663 (Resolved): Hadoop: file ownership/permission not available in hadoop
- This is still a pretty cheap fix :), but I think it's enough to close out this bug.
- 04:12 PM CephFS Bug #1663: Hadoop: file ownership/permission not available in hadoop
- a79b7e17ebbc70cedae80216986ae5fd52a1c0b7 provides an OK fix for now. Basically it makes any file look like the curren...
- 04:08 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Bummer. Well... for the time being it may be sufficient to force FileStatus.getModificationTime() to go directly to t...
- 03:58 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Yeah, it's not impossible, I just would have thought that one of the other updates would have prompted the server to ...
- 03:52 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Do you mean that you are surprised that client-1's inode didn't get updated from the server's change before the stat ...
- 03:49 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- If that's the case then I'm surprised the mtime didn't get updated at an earlier time. If nothing else we can probabl...
- 03:44 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Greg Farnum wrote:
> So the "bad" mtime is the same time the inode was created on the MDS server?
I think so. Her... - 03:35 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- So the "bad" mtime is the same time the inode was created on the MDS server?
- 03:30 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- If Client-1 is seeing a cached copy of the inode's mtime, then the following server-side scenario may explain what's ...
- 02:44 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Grepping for the inode number got me this:...
- 01:20 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Sage Weil wrote:
> If you can generate client logs for C1 and C2 (debug ms = 1, debug client = 10) that should tell ... - 11:44 AM CephFS Bug #1666: hadoop: time-related meta-data problems
- If you can generate client logs for C1 and C2 (debug ms = 1, debug client = 10) that should tell us everything.
- 11:07 AM CephFS Bug #1666: hadoop: time-related meta-data problems
- Just ran a little experiment that may shed some light on this....
- 03:49 PM CephFS Bug #1677: mds interval_set.h: 385: FAILED assert(p->first <= start)
- Here is the log from the MDS that caused this. I have from the other mds's, mon, and osd if it is relevant -- but not...
- 03:44 PM CephFS Bug #1677 (Resolved): mds interval_set.h: 385: FAILED assert(p->first <= start)
- Noah got this and sent it to the mailing list on Oct 28, 2011:...
- 02:15 PM Bug #1617 (New): pgs stuck down and peering with only one osd down and out
- Happened again today in teuthology:~teuthworker/archive/nightly_coverage_2011-11-03/1433:...
- 02:06 PM Messengers Bug #1674: daemons crash when sent random data
- This is actually going to be pretty unpleasant. Removing the asserts that deliberately crash on unexpected types is e...
- 06:29 AM Messengers Bug #1674 (Can't reproduce): daemons crash when sent random data
- mon seem to crash every time, osd seem to take a few attempts (similar stack trace). not tested mds...
- 12:04 PM Bug #1676 (Resolved): stats mismatch during snaps workunit
- It looks like this started failing between 10-20 and 10-24....
- 11:54 AM CephFS Bug #1675 (Can't reproduce): mds: failed rstat assert
- This happened during the multiple_rsync workunit.
From teuthology:~teuthworker/archive/nightly_coverage_2011-11-03/1... - 11:29 AM Bug #1671: rgw: access to swift auth url without user info crashes gateway
- Ah, failed to push. Rebased commit:0d4987d990e9795fda75d9e7903ba2d449b11fec.
- 02:52 AM Revision 376dad92 (ceph): hadoop: remove unused fs_default_name
- The variable fs_default_name is effectively unused
and the same affect is achieved by treating paths
in a standard wa... - 02:51 AM Revision 3191e0db (ceph): hadoop: FileSystem.rename should not return FileNotFound
- This fixes several unit test failure cases.
Signed-off-by: Noah Watkins <noahwatkins@gmail.com> - 02:51 AM Revision 60e1e148 (ceph): hadoop: ENOTDIR should be negative
- Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
- 02:51 AM Revision 6deea1c2 (ceph): hadoop: fix unit test: testWorkingDirectory
- The working directory should be set in initialize() and
is expected by the unit tests to be fully qualified (i.e.
wit... - 02:51 AM Revision ccb08e21 (ceph): hadoop: remove deprecation warning
- The routine cannot be fully removed yet because it
still exists as an abstract function in FileSystem class.
Signed-... - 02:51 AM Revision 1c24fc7a (ceph): hadoop: remove deprecated isDirectory()
- Uses the suggested getFileStatus() method for
replacing the deprecated isDirectory(). This is
only marginally slower ... - 02:51 AM Revision a407da0e (ceph): hadoop: remove statistics initialization
- This is already handled by super.initialize()
Signed-off-by: Noah Watkins <noahwatkins@gmail.com> - 02:51 AM Revision dcf2d629 (ceph): hadoop: remove unused variable
- Remove CephFileSystem.debug as log4j is now
used for debug level control.
Signed-off-by: Noah Watkins <noahwatkins@g... - 02:51 AM Revision 9e8fa029 (ceph): hadoop: remove initialization check
- The initialization check is removed because
it is part of Hadoop's treatment of file systems
that initialize() is cal... - 02:51 AM Revision 3006c6e5 (ceph): hadoop: simplify workingDir handling; add home directory
- 1. Simplifies the handling of paths by allowing them to be passed
around and manipulated in their fully qualified for... - 02:50 AM Revision a79b7e17 (ceph): hadoop: emulate Ceph file owner as current user
- Make CephFileSystem tell Hadoop that the owner
of all files is the current user. This provides
zero security or isola... - 02:49 AM Revision e9adf735 (ceph): hadoop: use standard log4j logging facility
- Replace ceph.debug(msg, level) with LOG.level(msg)
provided by the log4j facility used by Hadoop. The
level can now b... - 02:06 AM Bug #1529: cosd: os/FileStore.cc: 2390: FAILED assert(0 == "ENOENT on clone suggests osd bug")
- Sorry for the slow response! Somehow I didn't get a e-mail update.
I do have logs preceeding the crash, but they a...
11/02/2011
- 08:45 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Something like this would make the most sense to me. (I'd have to check the specifics of mtime updating to see exactly.)
- 08:30 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Formatting oops:...
- 08:29 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- You're right about that last point Greg, it doesn't quite add up--not thinking straight today.
Here is what happen... - 07:46 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- I'd have to look at the specifics again -- but it probably can't be done. If the client buffers a write and then flus...
- 06:39 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- So, I think I've got this nailed down. The good news is that the error was a clock sync issue. The bad news is that i...
- 06:51 PM Revision c861ee10 (ceph): PG: mark scrubmap entry as not absent when we see an update
- Previously, there would be an assert failure in _scan_list if we see an
object deleted and then recreated.
Signed-of... - 06:33 PM Revision a2f406ef (ceph): testrados: set CEPH_CLIENT_ID without a ;
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 04:23 PM Bug #1633: osd crash in CryptoKey::decrypt
- Happened again today. I put the core and tarball on the gcov gitbuilder in ~ubuntu/bug_1633.
- 03:45 PM Revision 78111d07 (ceph): Merge branch 'wip-freebsd'
- Conflicts:
src/osd/OSD.cc - 03:44 PM Revision 47b70367 (ceph): debian: update VCS sources
- Signed-off-by: Laszlo Boszormenyi <gcs@debian.hu>
- 03:44 PM Revision 0b0f65a4 (ceph): add missingok to logrotate
- When ceph is not running, it has no logs. Thus logrotate has nothing to
rotate. The missingok directive handles this ... - 03:44 PM Revision f4971328 (ceph): debian: empty dependency_libs in *.la files
- Per policy and multiarch support.
Signed-off-by: Laszlo Boszormenyi <gcs@debian.hu> - 03:44 PM Revision 26787ce3 (ceph): debian: add watch
- Signed-off-by: Laszlo Boszormenyi <gcs@debian.hu>
- 03:44 PM Revision ee34e09c (ceph): debian: fix libceph1 -> libcephfs1 rename
- Signed-off-by: Laszlo Boszormenyi <gcs@debian.hu>
- 02:34 PM rgw Bug #1673 (Won't Fix): rgw: mod_fastcgi needs to be backward compatible
- The changes we introduced for 100-continue breaks the protocol, we need to make that optional that way or another.
- 02:18 PM rgw Bug #1672 (Resolved): rgw: support chunked transfer encoding
- This is required for swift support. Currently mod_fastcgi doesn't support chunked transfer and we can't just use mod_...
- 01:33 PM Bug #1530: osd crash during build_inc_scrub_map
- Alright, in irc, slb seems to have hit a related bug...with logging!
- 11:49 AM Bug #1530: osd crash during build_inc_scrub_map
- c861ee105475b3f20f64f51b8611f9b69207ca8c should take care of the assert(!o.negative) error. Still trying to reproduc...
- 09:02 AM Bug #1530: osd crash during build_inc_scrub_map
- Possibly related: the snaps workunit failed yesterday and today with bad stats:...
- 08:53 AM Bug #1530: osd crash during build_inc_scrub_map
- Two more tests hit this last night, and two other osds crashed due to an assert in build_inc_scrub_map:...
- 12:41 PM Bug #1671 (Resolved): rgw: access to swift auth url without user info crashes gateway
- Fixed, commit:add8f59df9b6ef63a8431d3415e791b14ce1fe3c.
- 12:36 PM Bug #1671 (Resolved): rgw: access to swift auth url without user info crashes gateway
- 11:31 AM Bug #1657 (Resolved): teuthology: testrados failed to find conf
- Forgot to include my fix for that, pushed: a2f406ef49a1e5ec31d90957122e14addf56901c.
- 08:58 AM Bug #1657 (New): teuthology: testrados failed to find conf
- Failed due to escaped env setting:...
- 09:35 AM Bug #1670 (Can't reproduce): osd: crash in update_heartbeat_peers
- ...
- 04:20 AM Revision 2fc01b52 (ceph): osdmaptool: test --create-with-conf with racks
- Make sure we generate a map that will map (and not assert about bad
max_osd/max_device mismatch).
Signed-off-by: Sag... - 04:14 AM Revision 885d7148 (ceph): osdmap: assert that osdmap max_osds >= crushmap max_devices
- This will catch potential array overruns before they happen.
Signed-off-by: Sage Weil <sage@newdream.net> - 04:14 AM Revision 0bcdd4f3 (ceph): osdmap: fix off-by-one in build_simple_from_conf
- maxosd is the highest osd id. set_max_osd(that + 1), since that is
setting the array size. This fixes references of... - 03:04 AM Revision b66847ea (ceph): osd: fix assert include
- Signed-off-by: Sage Weil <sage@newdream.net>
11/01/2011
- 11:07 PM Bug #1669: linux 32 bit kernel client ld libraries and rm issue
- Yes it is much better. I used a git version of the kernel and it's version is 3.1.0+. It seems ldconfig and rm are ...
- 08:59 PM Bug #1669: linux 32 bit kernel client ld libraries and rm issue
- There was a recent fix for 32-bit ino generation that will avoid this problem most of the time, although in theory yo...
- 07:00 PM Bug #1669 (Resolved): linux 32 bit kernel client ld libraries and rm issue
- I am running ceph on 64 bit OS (Debian linux-image-3.0.0-2-x86_64). It is on two machines each of them having 1 mon,...
- 11:02 PM Revision 219141e9 (ceph): rgw: swift prefix and path params fixes
- 08:12 PM Revision 143c572b (ceph): .gitignore: test_str_list
- 08:10 PM Revision aa5f697f (ceph): Makefile: include/compat.h in tarball
- Signed-off-by: Sage Weil <sage@newdream.net>
- 07:35 PM Revision 9252dccc (ceph): Merge branch 'master' into wip-freebsd
- 06:49 PM Revision b3b45bf9 (ceph): Merge remote-tracking branch 'gh/wip-auth'
- 06:43 PM Revision 79d9718d (ceph): common: make get_str_list work with other delimiters, and skip the
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:43 PM Revision 99bcd7b5 (ceph): common: get_str_list unit tests
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:19 PM Revision ba8c345b (ceph): monclient: fail fast when our auth protocols aren't supported
- This handles the case where the server does not support any of the
authentication protocols that the client does. Pre... - 06:19 PM Revision 7a4c232f (ceph): monclient: fix else formatting
- If one branch has braces, the other should too.
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> - 06:16 PM Revision d1e95134 (ceph): PG: set_last_peering_reset in Reset constructor
- If an osd in the prior set comes up, we can restart peering without a
new peering interval starting. However, we sti... - 05:46 PM Revision e15177ab (ceph): monclient: fail fast when our auth protocols aren't supported
- This handles the case where the server does not support any of the
authentication protocols that the client does. Pre... - 05:46 PM Revision ef51f0fa (ceph): monclient: fix else formatting
- If one branch has braces, the other should too.
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> - 02:56 PM Bug #1633: osd crash in CryptoKey::decrypt
- have a core but no matching binary :(. need to reproduce again, and save the build tarball.
- 01:03 PM devops Feature #1668 (New): collectd: push ceph plugin upstream
- Rebase the perfcounter ceph plugin in the dho collectd repo against mainline collectd and push upstream.
- 11:09 AM Bug #1530: osd crash during build_inc_scrub_map
- can someone work on reproducing this? see metropolis:~sage/src/teuthology/j.1530 and hammer.sh
- 10:11 AM Bug #1530: osd crash during build_inc_scrub_map
- This happened again in teuthology:~teuthworker/archive/nightly_coverage_2011-11-01/1254/remote/ubuntu@sepia68.ceph.dr...
- 11:08 AM CephFS Bug #1549: mds: zeroed root CDir* vtable in scatter_writebehind_finish
- Someone needs to try to reproduce this with logs. fwiw metropolis:~sage/src/teuthology/hammer.sh is what i've been u...
- 10:22 AM CephFS Bug #1549: mds: zeroed root CDir* vtable in scatter_writebehind_finish
- This happened after the misc workunit today.
- 08:49 AM Linux kernel client Bug #1667 (Resolved): BUG at fs/inode.c line 1375
- ...
10/31/2011
- 10:03 PM Revision 9ea02239 (ceph): osd: kill unused on_osd_failure() hook
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:00 PM Revision 1d9e8065 (ceph): RadosModel.h: use default conf location
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 09:54 PM Revision 810cae1a (ceph): testrados: specify CEPH_CONF directly
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 09:02 PM Revision b9a0b2b7 (ceph): Revert "PG: call set_last_peering_reset in Started contructor"
- Unfortunately, the Started constructor doesn't occur until map
activation. We need to reset last_peering_reset exact... - 06:15 PM Revision f9b7ecdb (ceph): hadoop: Return NULL when the path does not exist.
- Although unspecified in the declaration header, other file
systems return a single result when the path is a file.
T... - 05:53 PM Bug #1633: osd crash in CryptoKey::decrypt
- Another occurrence in teuthology:~teuthology/archive/nightly_coverage_2011-10-28/1170/remote/ubuntu@sepia50.ceph.drea...
- 05:32 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- It looks like the check is equality of timestamps. So, I think Hadoop is setting an explicit timestamp, and sometime ...
- 05:30 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- All of the local clocks on the nodes look good. The code is comparing timestamps (I assume since epoch), so maybe the...
- 05:06 PM CephFS Bug #1666: hadoop: time-related meta-data problems
- Neither of these errors are in code that's remotely familiar to me. So my first favorite question is:
Are your clock... - 04:55 PM CephFS Bug #1666 (Resolved): hadoop: time-related meta-data problems
- The following exceptions are being thrown. It looks like something related to lstat?
pre>
java.io.IOException: Th... - 02:59 PM Bug #1657 (Resolved): teuthology: testrados failed to find conf
- Should work now
ceph: 1d9e8065c835c343608930585c2853984cde2fa8
teuthology: 810cae1a1d03138abfa54cd31059723ec0c22ab1 - 02:04 PM Bug #1665 (Resolved): osd: last_peering_reset incorrect on stray?
- b9a0b2b7a4d3b5a7db1f942af0158712199377a8 reverted 6d123067ce1ba99522281d5c72623bd5ba3e0fc8
- 12:09 PM Bug #1665: osd: last_peering_reset incorrect on stray?
- this is why. the interval starts at 150,a nd that is when teh query is sent. on the stray, we hit it in 151:...
- 11:46 AM Bug #1665 (Resolved): osd: last_peering_reset incorrect on stray?
- on alexandria,...
- 01:55 PM Bug #1588 (Can't reproduce): blogbench on kclient possibly made machine die
- I think this is fixed - the nightly tests haven't hit it in the past week, since 339573406737461cfb17bebabf7ba536a302...
- 11:35 AM CephFS Bug #1661 (Resolved): Hadoop: expected system directories not present
- Apparently this was actually the result of an API mismatch. Fixed by Noah's patch in commit:f9b7ecdb5bba1439dc4c13005...
- 11:26 AM Feature #1618: libvirt: make sure migration works
- Braindump of what I did for the earlier libvirt migration demo:
- on each vm host, install kvm 0.15 (0.14 is too o... - 09:13 AM Bug #1415 (Duplicate): cosd assertion: existing->state == STATE_CONNECTING || existing->state ==...
- 09:11 AM rgw Feature #1664 (Resolved): rgw: pass swift tests
- 09:06 AM Messengers Feature #1648 (Duplicate): msgr: choose ip to bind to based on network
- 09:02 AM Messengers Feature #1648: msgr: choose ip to bind to based on network
- duplicates #1487
- 07:58 AM Bug #1529: cosd: os/FileStore.cc: 2390: FAILED assert(0 == "ENOENT on clone suggests osd bug")
- Sage Weil wrote:
> Do you have the odd log preferring the restart?
Er, osd log preceeding ... - 07:54 AM Bug #1529: cosd: os/FileStore.cc: 2390: FAILED assert(0 == "ENOENT on clone suggests osd bug")
- Do you have the odd log preferring the restart?
- 06:46 AM Bug #1529: cosd: os/FileStore.cc: 2390: FAILED assert(0 == "ENOENT on clone suggests osd bug")
- I'm still seeing this one. All my 6 OSDs went down and after starting them most of them would crash:...
10/30/2011
- 12:42 AM Revision 5bd029ef (ceph): osdmap: fix g_ceph_context reference
- Use cct.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
10/28/2011
- 10:48 PM Revision 0fa86182 (ceph): ReplicatedPG: check for peering restart before share_pg_info
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 09:33 PM Revision 199e04ab (ceph): mkcephfs: build initial osdmap from information in ceph.conf
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:32 PM Revision 3f678931 (ceph): crush: make insert_item take float for weight
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:32 PM Revision 07c9de83 (ceph): osdmaptool: build initial map from ceph.conf
- This builds the intial osd and crush maps from what is in the ceph.conf,
taking advantage of host or rack tags that a... - 09:25 PM Revision ef4b95c8 (ceph): ReplicatedPG: Clean up old snap links when recovering a clone
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 09:25 PM Revision bd3223f9 (ceph): PG: Create new snap directories independently on replica
- Previously, we shipped over the collection creation as part
of the transaction. However, the snap directory on the
r... - 09:04 PM Revision b497b385 (ceph): rgw: canonical resource should use unencoded url
- 08:00 PM Revision 5fe8e00a (ceph): Merge pull request #4 from vzctl/master
- fix error: 'snprintf' was not declared in this scope
- 06:49 PM Revision a8450005 (ceph): rgw: cleanup, remove unused user_id
- Some access methods required user_id param, but that was never really used. At
this point we should just remove them. - 06:42 PM Revision 7ee0747c (ceph): mkcephfs: skip non-btrfs osds even with --mkbtrfs
- This lets you do a mixed btrfs and non-btrfs file systems.
Signed-off-by: Sage Weil <sage@newdream.net> - 05:39 PM Revision 2bb283ba (ceph): Merge branch 'stable'
- 05:38 PM Revision 3a17f023 (ceph): debian: break redundant dependencies
- They confuse APT it seems.
ceph-common -> librbd1 -> librados2
radosgw -> ceph-common -> librados2
Signed-off-by:... - 05:05 PM Revision b8979f4d (ceph): MOSDMap: do not leave {oldest,newest}_map uninitialized when decoding o...
- This leads to badness like
osd_map(295..296 src has 74308224..0) v1
Signed-off-by: Sage Weil <sage@newdream.net> - 03:46 PM CephFS Bug #1661: Hadoop: expected system directories not present
- Blindly creating directories is definitely not the proper solution. Somebody will need to take the time to figure out...
- 03:32 PM CephFS Bug #1661: Hadoop: expected system directories not present
- In this particular instance it is a map-reduce specific directory. I suspect that MapReduce is responsible for this, ...
- 03:22 PM CephFS Bug #1661: Hadoop: expected system directories not present
- Sounds to me like CephFileSystem should just create the directory if it doesn't exist..
- 03:13 PM CephFS Bug #1661: Hadoop: expected system directories not present
- Good to know. I think at this point I need to paper over many things, but want to record all these issues. I'll just ...
- 03:08 PM CephFS Bug #1661: Hadoop: expected system directories not present
- I remember running into this issue when developing things and deciding to just paper over it at the time -- I couldn'...
- 03:05 PM CephFS Bug #1661: Hadoop: expected system directories not present
- Adding: when this directory is created by hand before map reduce starts the error is gone.
- 03:04 PM CephFS Bug #1661 (Resolved): Hadoop: expected system directories not present
- Hadoop complains that directories within the file system that are expected to be present are not present. Hadoop may ...
- 03:24 PM CephFS Bug #1663: Hadoop: file ownership/permission not available in hadoop
- Noah Watkins wrote:
> This is a very simple hack that will make hadoop ignore the permission for the time being:
... - 03:23 PM CephFS Bug #1663: Hadoop: file ownership/permission not available in hadoop
- This is a very simple hack that will make hadoop ignore the permission for the time being:
diff --git a/src/mapred... - 03:16 PM CephFS Bug #1663 (Resolved): Hadoop: file ownership/permission not available in hadoop
- Hadoop complains about incorrect file ownership. An 'ls' via Hadoop FS interface reveals no permission information, b...
- 03:08 PM rbd Feature #1662 (Resolved): libvirt: obscure qemu/rbd secrets
- 02:36 PM Feature #1067 (Resolved): mkcephfs: magically group osds on same host into subtrees in the genera...
- commit:199e04aba1bd3d0c5a2a0e13e4500bef9cc206cf
- 01:46 PM Revision 6353d7b5 (ceph): include stdio in order to fix snprintf compilation error
- Signed-off-by: Alexey Lapitsky <lex@realisticgroup.com>
- 12:09 PM rgw Feature #1660 (New): add a teuthology test to verify rgw log summation
- 12:08 PM rgw Bug #1645 (Resolved): rgw bucket suspended broken
- Fixed, commit:6752babdfda1be0524d82b84adfa4663aded32f6. Also added a teuthology test.
- 09:30 AM rgw Feature #829 (Resolved): rgw: support swift POST
- We actually support now swift POST for metadata changes. For ACL changes there's issue #830.
- 09:28 AM rgw Bug #1643: radosgw-admin log show should accept --time
- The problem is that the logs are indexed by date, and not by time. Filtering by time means that we need to scan the o...
- 04:04 AM Revision 46bb82f5 (ceph): client: fix return value for _readdir_cache_cb
- Return 0 for end of directory here, too.
Clarify some comments.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:28 AM Revision 943893e8 (ceph): ceph: fix snprintf warning
- warning: tools/ceph.cc:146: format not a string literal and no format arguments
Signed-off-by: Sage Weil <sage.weil@... - 01:12 AM Revision 64992113 (ceph): auth: return unknown if no supported auth is found
- If NONE is supported, it will already be in the list of supported
protocols, so there's no need to default to it here... - 01:01 AM Bug #1659 (Can't reproduce): Upgrade from 0.27 -> 0.37 going wrong, OSDs miss map updates
- Hi,
Like I mentioned on IRC, I had some problems with upgrading my cluster from 0.27 to 0.37.
It was a big step... - 12:24 AM Revision 1a4eec20 (ceph): uclient: fix _getdents and add some documentation.
- If readdir_r_cb returns 0, that means SUCCESS, regardless of how
many entries it actually wrote.
If it returns <0, th...
10/27/2011
- 11:15 PM Revision 27ec04e7 (ceph): cfuse: remove unneeded loop.
- The only time this was looping previously was completely unnecessary
anyway, as 1 meant the same thing as 0: there ar... - 11:15 PM Revision e37ab416 (ceph): uclient: align readdirplus_r with readdir_r.
- The only user of this code expects to get 1 on a successfully-filled
value, 0 on a successful non-fill, or -errno oth... - 11:15 PM Revision 55aace73 (ceph): uclient: readdir_r_cb documentation, and it only returns 0 or -errno.
- Returning 0 or 1 in different situations that were effectively the
same is useless and confusing.
Signed-off-by: Gre... - 09:35 PM Revision 354055f8 (ceph): rgw: swift related adjustments
- 09:26 PM Revision 713a4428 (ceph): Merge branch 'master' of github.com:NewDreamNetwork/ceph
- 09:04 PM Revision ed839f5a (ceph): fixed graphic reference and headings
- 09:00 PM Revision 2c4eb075 (ceph): fixed image reference
- 08:54 PM Revision b42443ec (ceph): fixed architecture document
- 08:43 PM Revision c57ed06c (ceph): add images for documentation
- 07:51 PM Revision 7a022029 (ceph): rgw: handle swift PUT with incorrect etag
- 07:44 PM Revision cae7d5a0 (ceph): rgw: handle swift PUT with incorrect etag
- 07:44 PM Revision 697bba39 (ceph): rgw: handle swift PUT with incorrect etag
- 07:11 PM Revision 10c35087 (ceph): rgw: add user suspend/enable test
- 06:32 PM Revision 86aa940f (ceph): rgw: log-to-stderr is now a binary flag
- 06:20 PM Revision a817a38e (ceph): rgw: handle swift PUT with incorrect etag
- 06:16 PM Revision d9dfd147 (ceph): rgw: handle swift PUT with incorrect etag
- 06:02 PM Revision 87224c08 (ceph): rgw: handle swift PUT with incorrect etag
- 05:02 PM Revision e4dcbd03 (ceph): ceph: refactor for generic --admin-daemon <sock> <cmd> too
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:50 PM Revision 6979eaa0 (ceph): filejournal: journal_replay_from
- Force journal replay from a point other than the op_seq recorded by the
fs. This is useful if you want to skip bad e... - 04:50 PM Revision 89dccc0e (ceph): ceph: --dump-perf-counters[-schema] sockpath
- Quick and dirty way to dump perfcounters stats. Not documenting this until
we decide this is where it should live.
... - 04:26 PM Revision a9b75f21 (ceph): Merge branch 'stable'
- 04:26 PM Revision b3e1e3e1 (ceph): rados: improve error message
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:46 AM CephFS Bug #1549 (Need More Info): mds: zeroed root CDir* vtable in scatter_writebehind_finish
- bleh. need logs... i'll start this up in a loop again.
- 10:33 AM CephFS Bug #1549: mds: zeroed root CDir* vtable in scatter_writebehind_finish
- This happened again today after fsstress. From teuthology:~teuthworker/archive/nightly_coverage_2011-10-27/1083/teuth...
- 09:26 AM Feature #1658 (Resolved): osd: backfill instead of backlog
- 08:59 AM Feature #1646: mon: catch up on committed items before attempting to join quorum
- Not sure exactly what you mean, but that sounds a bit like the behavior when the encoding changes and the monitors ar...
- 03:55 AM Feature #1646: mon: catch up on committed items before attempting to join quorum
- Any chance this is related with an issue I noticed last night, in which the primary mon was receiving and displaying ...
- 04:20 AM Revision 11691a71 (ceph): radosgw-admin: fix key create check
- Also fixes warning
warning: rgw/rgw_admin.cc:812: suggest parentheses around ‘&&’ within ‘||’
Signed-off-by: Sage W... - 12:24 AM Revision 921ce53d (ceph): osd: guard checks for writes
- fa722de6708d3e92037df6289cc29ece12c8ea66 moved these checks, and
accidentally removed the may_write() guard. This cau... - 12:20 AM Revision 0c78f0dc (ceph): rgw: handle swift PUT with incorrect etag
- 12:00 AM Revision 213eb13d (ceph): Revert "hadoop: get hadoop bindings to build again" and fix.
- It's just wrong. The Java code is still passing a String along
regardless of what you ask the C to do! Fix it by grab...
10/26/2011
- 11:07 PM Revision e8e10158 (ceph): rgw: rgw-admin --skip-zero-entries
- 11:00 PM Revision 180c744b (ceph): perfcounters: fix accessor name
- FreakingCamelCaps
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 11:00 PM Revision 1a0a732e (ceph): objecter: instrument with perfcounter
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:34 PM Revision e747456c (ceph): rgw: rgw-admin generate-key/access-key=false fix
- 10:34 PM Revision 9386a7b5 (ceph): rgw: rgw-admin can show log summation
- 09:56 PM Revision 7fbf28a9 (ceph): osd: read_log: only list the collection once
- After upgrading we may need to list the collection to recover the hash
value when upgrading an old collection.
Signe... - 09:30 PM Revision 6752babd (ceph): rgw: fix bucket suspension
- 05:46 PM Bug #1654 (Resolved): snaps workunit failed on cfuse
- Fixed by 921ce53d6efc3f1bf7056f05467aff5c3104dcc8.
- 03:24 PM Bug #1654: snaps workunit failed on cfuse
- And the librados selfmanaged snaps tests also failed with an unexpected EINVAL when reading from a snapshot.
- 11:39 AM Bug #1654: snaps workunit failed on cfuse
- There might have been a bug introduced in snapshot contexts - two rbd tests got EINVAL when setting a snapshot, meani...
- 11:35 AM Bug #1654 (Resolved): snaps workunit failed on cfuse
- ...
- 05:31 PM Bug #1657 (Resolved): teuthology: testrados failed to find conf
- From teuthology:~teuthworker/archive/nightly_coverage_2011-10-26/1037/teuthology.log:...
- 04:11 PM rgw Feature #773: rgw: efficient list-objects filtering
- With the new osd-class index, this should be pretty straight forward.
- 04:09 PM rgw Feature #1641 (Rejected): radosgw-admn log show --bandwidth-only
- commit:9386a7b5e57de4994ff3ad4987ef309cb8275392 added data aggregation, so there's no need to dump the entire log now...
- 04:06 PM rgw Feature #1642 (Resolved): radosgw-admin log show --nonzero-only
- Fixed, commit:e8e101580ea04628713f51171e9af58aec1acbd2.
rgw-admin accepts --skip-zero-entries now. - 04:03 PM CephFS Bug #1656: Hadoop client unit test failures
- Sounds good to me -- which patches we want to keep in the tree are probably a management decision but I'm happy to pu...
- 03:55 PM CephFS Bug #1656: Hadoop client unit test failures
- Alright, so I think at this point I'd like to see two patches:
1) A patch against the downloadable tarball (much e... - 03:49 PM CephFS Bug #1656: Hadoop client unit test failures
- I believe the patch was made against the then-current svn 0.21 branch (which is now very dead). I pushed changes to t...
- 03:39 PM CephFS Bug #1656: Hadoop client unit test failures
- This was hadoop-0.20.205.0 with the latest Ceph master branch.
It looked like the patch in src/client/hadoop was o... - 03:30 PM CephFS Bug #1656: Hadoop client unit test failures
- What versions of the systems were you running when these failed?
I don't remember how they're set up but they migh... - 01:59 PM CephFS Bug #1656 (Won't Fix): Hadoop client unit test failures
- The Ceph Hadoop File System passes nearly all its tests except a few. I've included the test log below that shows the...
- 03:38 PM Bug #1555 (Resolved): radosgw_admin --gen-access-key=false and --gen-secret=false flags appear to...
- Fixed, commit:e747456c9f6cc8cc0367bb80e757b1b24e098de1.
- 01:49 PM Feature #1655 (Resolved): gitbuilder aggregator page
- single page that has 1 line per gitbuilder, with instance name and then the top line of the gitbuilder status screen ...
- 10:13 AM Bug #1590 (Duplicate): occasionally excessive mon memory footprint
- 10:12 AM Bug #1590: occasionally excessive mon memory footprint
- this will go away with #1646.
- 10:11 AM Bug #1634 (Can't reproduce): osd: crash decoding non-existent object_info_t
- going to see if this comes up again after this last round of osd fixes
- 09:58 AM Feature #1653 (Resolved): librados: python binding nose tests
- 04:34 AM Revision f197e845 (ceph): rgw: fix uninitialized variable warnings
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
10/25/2011
- 11:39 PM Revision 952be11a (ceph): hadoop: bring back Java changes.
- These convert the Hadoop stuff to work on the branch-0.20 API.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.... - 11:29 PM Revision 71fd8302 (ceph): Merge branch 'master' of ssh://github.com/NewDreamNetwork/ceph
- Conflicts:
src/rgw/rgw_rados.cc - 11:23 PM Revision d9f73605 (ceph): rgw: fix attr cache
- 10:35 PM Bug #1628 (Resolved): segfault attempting to map an rbd snapshot
- 10:33 PM Bug #1099 (Closed): osd: handle recovery of lost objects
- this has been reimplemented (at least the revert case).
- 10:32 PM Cleanup #146 (Rejected): Complete build options for Pthread API
- 10:29 PM Feature #641 (Rejected): allow logs to be piped to an external program
- works for me.
- 10:28 PM Bug #250 (Resolved): mon: delete old states to avoid filling disk
- 10:28 PM Feature #875 (Resolved): osd: clean up old osdmaps
- 10:24 PM Feature #1649 (Resolved): osd: make replay interval a per-pool setting
- 10:08 PM Revision 5151a8af (ceph): common/ceph_extattr.[ch] > common/xattr.[ch]
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:54 PM Revision 46f330d0 (ceph): Merge branch 'master' into wip-freebsd
- 09:15 PM Revision ef48183a (ceph): fix osdmaptool clitests
- 09:02 PM Revision 8ae02dab (ceph): Merge branch 'wip-pools'
- 05:52 PM Revision 6287ccf6 (ceph): mon: reencode routed messages
- The message encoding may depend on the target features. Clear the
payload so that the Message gets reencoded appropr... - 05:51 PM Revision 72e0ca02 (ceph): MOSDMap: reencode full map embedded in Incremental, as needed
- The Incremental may have a bufferlist containing a full map; reencode
that too if we are reencoding for old clients.
... - 05:13 PM Revision cd6d7009 (ceph): Merge remote-tracking branch 'gh/wip-rbd-tool'
- 04:53 PM Revision 6ca99060 (ceph): mon: parse 0 values properly
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:53 PM Revision 90f0429f (ceph): mon: fix rare races with pool updates
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:48 PM CephFS Bug #1114 (Need More Info): NFS export extreme slowdown
- Need to reproduce this on the current trunk and fully characterize what is going on.
- the the nfs server in sync ... - 04:46 PM Bug #1194 (Resolved): kclient: NFS reexport does not survive ceph fs remount
- going to assume the above fixed it until we hear otherwise :)
- 03:50 PM CephFS Bug #1585 (Can't reproduce): mds crash during shutdown
- 03:38 PM Bug #1629 (Can't reproduce): pgs stuck degraded (only mapped to 1 osd)
- pre-prior set refactor and current round of thrashing fixes.
- 03:34 PM Bug #1624 (Resolved): osd crash in HearbeatMap::_check
- going to chalk these up to the infinite loop fixed in that previous patch.
- 03:33 PM Bug #1617 (Rejected): pgs stuck down and peering with only one osd down and out
- non-specific, and pre-prior set refactor.
- 03:31 PM Bug #1311 (Closed): qa: TestSnaps: stuck in active
- ancient and presumably covered by current thrashing tests
- 03:30 PM Bug #1292 (Closed): qa: bench & thrashosd PG won't go clean
- this is ancient and presumably covered by the new thrashing tests.
- 03:29 PM Bug #1609 (Resolved): osd: failed assert(info.last_complete == info.last_update)
- lots of stuff, mainly commit:03ad5a28eee2328eb2419c48a14df1a3624fc4c7
- 10:31 AM Bug #1526 (Resolved): log bound mismatch after thrashing with bonnie
- 05:51 AM Revision 43aa33a2 (ceph): Merge remote branch 'gh/wip-osd-queue'
- 05:50 AM Revision 7de2f7a9 (ceph): osd: print useful debug info from choose_acting
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:50 AM Revision c30ab1e2 (ceph): osd: MOSDPGNotify: print prettier
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:50 AM Revision 12b3b2d5 (ceph): osd: fix generate_past_intervals maybe_went_rw on oldest interval
- We stop working backwards when we hit last_epoch_clean, which means for the
oldest interval first_epoch may not be th... - 05:50 AM Revision 03ad5a28 (ceph): osd: fix last_complete adjustment after recovering an object
- After we recover each object, we try to raise the last_complete value
(and matching complete_to iterator). If our lo... - 05:50 AM Revision e2f3c20b (ceph): osd: make proc_replica_log missing dump include useful information
- I needed to see have/need to debug a weird unfound issue turned up by
thrashing.
Signed-off-by: Sage Weil <sage@newd... - 05:21 AM Revision f8e92896 (ceph): osd: fix/simplify op discard checks
- Use a helper to determine when we should discard an op due to the client
being disconnected. Use this when the op is... - 05:13 AM Revision fa722de6 (ceph): osd: move queue checks into enqueue_op, kill _handle_ helpers
- This simplifies things, and renames the checks to make it clear that we are
doing validation checks only, with no sid... - 04:59 AM Revision 3a2dc656 (ceph): osd: move op cap check into helper
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:54 AM Revision b17c9ca5 (ceph): osd: handle missing/degraded in op thread
- The _handle_op() method (and friends) are called when an op is initially
queued and when it is requeued. In the requ... - 04:54 AM Revision b1de9131 (ceph): osd: drop ability to disable op queue entirely
- This is pretty useless, and broken wrt requeueing anyway.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:54 AM Revision 662414d7 (ceph): osd: drop useless PG hooks
- These no longer need to be exposed to the generic OSD code.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:54 AM Revision 7aa0d89b (ceph): osd: set reqid on push/pull ops
- Not strictly necessary, but makes logs easier to follow.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:42 AM Revision e2766bd8 (ceph): mon: remove compatset cruft
- The CompatSet is built on demand; it's no longer static.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
10/24/2011
- 11:54 PM Revision 6f1b65c6 (ceph): ReplicatedPG: fix snapshot directory handling in snap_trimmer
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 11:54 PM Revision 024bcc4b (ceph): FileStore: ignore EEXIST on clones and collection creation !btrfs_snap
- We need to ignore EEXIST on btrfs also when m_filestore_btrfs_snap is
disabled.
Signed-off-by: Samuel Just <samuel.j... - 11:43 PM Revision 4d884040 (ceph): rgw: fix rgw_obj compare function
- 10:34 PM Revision df2967a6 (ceph): rgw: use a uint64_t instead of a size_t for storing the size
- librados uses uint64_t so that 32-bit architectures aren't hobbled.
Signed-off-by: Greg Farnum <gregory.farnum@dream... - 10:32 PM Revision 4b10cad8 (ceph): rbd: check command before opening the image
- Now map/unmap won't use librbd, and commands that don't take --snap
will give an error when it's used.
Signed-off-by... - 10:32 PM Revision 8c6db18d (ceph): rbd: specify which commands take --snap in usage
- Maybe this will be less confusing.
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> - 10:32 PM Revision 46bb4122 (ceph): rbd: let all commands use the pool/image@snapshot format
- This way you aren't forced to use '-p' or '--snap' to specify a pool
or snapshot for some commands.
Signed-off-by: J... - 10:32 PM Revision afa34794 (ceph): librbd: show correct size for snapshots
- header.size is the current size of the image.
ImageCtx::get_image_size() already does the right thing for
snapshots.
... - 10:32 PM Revision f4aa69a8 (ceph): workunit: check that rbd info returns the right size for snapshots
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:32 PM Revision e2296c3a (ceph): clitests: add rbd usage and invalid snap usage tests
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:32 PM Revision 93ccccd7 (ceph): rbd: remove unnecessary condition
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:32 PM Revision bfb5ceb2 (ceph): workunits: add rbd rollback and snapshot removal tests
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:32 PM Revision 315ab94e (ceph): librbd: propagate error from snap_set
- Previously rbd_snap_set always returned 0, even when the snapshot did
not exist.
Signed-off-by: Josh Durgin <josh.du... - 10:32 PM Revision a5a8a9cf (ceph): test_rbd: add a test for rolling back after resizing
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:32 PM Revision ae91911c (ceph): librbd: resize if necessary before rolling back
- This is a partial fix for test_rbd.TestImage.test_rollback_with_resize
Signed-off-by: Josh Durgin <josh.durgin@dream... - 10:32 PM Revision 2af32a41 (ceph): librados: use stored snap context for all operations
- Using an empty snap context led to the failure of
test_rbd.TestImage.test_rollback_with_resize, since clones weren't
... - 10:32 PM Revision b7aa57ff (ceph): rbd.py: update python bindings for new copy interface
- It was changed to return 0 on success in d7f7a213546b599d2eec4c6617593d232b43a7d6
Signed-off-by: Josh Durgin <josh.d... - 10:32 PM Revision e161ce15 (ceph): workunits: test rbd python bindings
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:15 PM Revision 2be3999d (ceph): Add btrfs dimension to thrash tasks
- Thrash tasks will now also run with and without btrfs.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 09:30 PM Revision 2ad6545a (ceph): Add testrados based thrashing tasks
- readwrite.yaml runs a read/write workload against a set of objects.
snaps.yaml adds snaps and rollback.
Signed-off-b... - 09:25 PM Revision 8d0a7c59 (ceph): testrados: rename testsnaps to testrados and make snap testing optional
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 08:52 PM Revision a1249d07 (ceph): workunit: set PYTHONPATH so we can test python bindings
- 06:46 PM Revision 88905b3a (ceph): test/osd: Add TestReadWrite
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 06:27 PM Revision 5e4e7972 (ceph): mon: allow adjustment of per-pool crash_replay_interval
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:12 PM Revision 40b7b572 (ceph): Merge branch 'rgw-dir-cleanup'
- 05:06 PM Revision f57c33df (ceph): rgw: fix check_disk_state; add a strip_namespace function.
- Use copies of the IoCtx rather than references so that
we can set locators without breaking stuff, and make use of th... - 05:04 PM Revision 0da45ca6 (ceph): rgw: rename translate_raw_obj to translate_raw_obj_to_obj_in_ns
- And document it. Because the naming is so bad that neither I nor
the author noticed it wasn't doing what we wanted it... - 05:04 PM Revision 927c3577 (ceph): rgw: add locators to the directory objects, and functions handling them
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 03:55 PM Linux kernel client Bug #1652 (Resolved): rbd: rollback correctly after resizing
- I just fixed this bug in librbd, but it seems the kernel has it too. If you take a snapshot, resize the image, then r...
- 11:38 AM rgw Bug #1567 (Resolved): rgw [list|delete]_bucket should clean up
- Merged into master in commit:40b7b57239515bd0794ef5da2477a2c5eb7a85e4.
Passed s3tests with a greatly-reduced timeo... - 10:53 AM Feature #1651 (Resolved): command line tool to interact with admin socket
- Maybe something like 'ceph --socket /var/run/ceph/osd.0.asok foo'?
- 04:07 AM Revision f37b08f8 (ceph): librados: behave if shutdown is called twice
- On failure, we shut ourselves down. If the caller calls shutdown again,
don't crash.
Fixes: #1650
Signed-off-by: Sa... - 04:05 AM Revision c15e62aa (ceph): mon: need to print pool id for output to be useful
- Signed-off-by: Sage Weil <sage@newdream.net>
- 03:40 AM Revision 8a087729 (ceph): mon: PGMap::dump: fix order in totals
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 02:01 AM Revision 1b941390 (ceph): osd: make osd dump slightly more concise
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:13 AM Revision 34c2f6a4 (ceph): osd: pg_pool_t: set crash_replay_interval on data pool when decoding old
- We want to preserve the crash_replay_interval on old clusters being
upgraded. Kludge this by setting it to 60 (the o...
10/23/2011
- 11:26 PM Revision 6779eb39 (ceph): osd: make osd replay interval a per-pool property
- Change the config value to only control the interval set when the data
pool is first created (presumably during mkfs)... - 11:26 PM Revision 8bb8e85d (ceph): Merge remote-tracking branch 'gh/master' into n
- Conflicts:
src/osd/OSDMap.h - 11:24 PM Revision f2816a1e (ceph): osd: pg_pool_t: normalize encoding
- Normalize encoding to be less awkward. Use a FEATURE bit to indicate
whether the new encoding is supported, and enco... - 11:24 PM Revision 7cb4d25d (ceph): osd: pg_pool_t: introduce flags, crash_replay_interval
- Introduce a per-pool crash_replay_interval so we can control whether
the OSD waits for replayed ACKed but not COMMITt... - 09:35 PM Bug #1650 (Resolved): “rados df” joins on thread never started with mons down or laggy (regressio...
- fixed by commit:f37b08f821a54263847e2c5c095bba5750908f86
- 07:56 PM Bug #1650 (Resolved): “rados df” joins on thread never started with mons down or laggy (regressio...
- If rados's attempt to connect the mons time out, it prints:
# rados df
couldn't connect to cluster! error -110
c... - 05:30 PM Revision 61cbb321 (ceph): ceph.conf: python parser doens't like ; comments
- 05:16 AM Revision 3ed06562 (ceph): ceph.conf: more frequent osd scrubbing; remove old cruft
- 03:44 AM Revision 54e28263 (ceph): scratchtool[pp]: fix rados_conf_set/get test of log_to_stderr
- Fix this warning
warning: scratchtool.c:142: comparison with string literal results in unspecified behavior
and fli... - 03:41 AM Revision 9323f25a (ceph): osd: fix PG::Log::copy_after wrt backlogs (again)
- Commit 68fe748fc2d703623050e8f2a448a0fd31ca8a0f fixed half of this problem,
but set this->tail incorrectly. If we re...
10/22/2011
- 10:13 PM Bug #1530: osd crash during build_inc_scrub_map
- I'm going to up the scrub frequency in the teuthology conf to help shake out these problems. There was another bug r...
- 10:07 PM Bug #1616 (Resolved): crash in is_supported_auth
- 10:06 PM Bug #1631 (Need More Info): osd: failed assert(repop_queue.front() == repop)
- need an osd log on this one
- 10:05 PM Cleanup #1644 (Resolved): osd: prior_set refactor
- 01:01 PM Bug #1471: osd: destroy_collection on non-empty dir
- I'm actually hitting the same bug with v0.37
It was time to upgrade my old (and good running!) 0.27 cluster to the...
10/21/2011
- 11:36 PM Revision 1b846f43 (ceph): radosgw: drop useless/broken set_val daemonize
- Not sure what the intent was here anyway... but it is broken (the func
takes a string, not a bool).
Signed-off-by: S... - 11:35 PM Revision 1f7cb757 (ceph): config: separate --log-to-stderr and --err-to-stderr
- Instead of having magic values (1 == errors only to stderr, 2 =
everything), have two booleans.
Signed-off-by: Sage ... - 11:14 PM Revision e98cbc43 (ceph): rgw: fix xattrs cache
- 10:24 PM Revision cf6a9404 (ceph): osd: eliminate CRASHED state
- This was an intermediate state that indicated that replay would be needed.
It was poorly named, and not very useful. ... - 10:24 PM Revision 03593019 (ceph): mon: fix last_clean_interval calculation
- This up_rom == first check is old and wrong. It may have been correct at
the time, when the OSD had a defined shutdo... - 10:24 PM Revision 600bda47 (ceph): osd: fix last_clean interval bounds
- It was _first and _last, inclusive, but the epochs are really points in
time, so _last should have been non-inclusive... - 10:24 PM Revision 249ed569 (ceph): osd: move may_need_replay calculation out of PriorSet
- Although they both depend on past intervals, they are unrelated. Factor
out the may_need_replay calculation from Pri... - 10:24 PM Revision 30c34ab8 (ceph): osd: trim past intervals when we complete recovery.
- We weren't trimming at all, which meant these would just accumulate
indefinitely.
Signed-off-by: Sage Weil <sage@new... - 10:14 PM Revision d6661f93 (ceph): ReplicatedPG: Include pg version in MOSDOpReply on error
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 08:23 PM Revision f8afd8bf (ceph): rgw: reduce rados bucket stats (and getxattrs)
- we didn't pass the context, and some other issue with the context map
- 05:54 PM Revision b8beff3d (ceph): ceph_manager: count active+clean+<somjething else> as active+clean
- In my case, one pg was active+clean+scrubbing.
Signed-off-by: Sage Weil <sage@newdream.net> - 05:32 PM Revision a1756c5e (ceph): rgw: object removal should remove object from index anyway
- even if object doesn't exist. Index might have the wrong info.
- 04:56 PM Revision dd5087fa (ceph): osd: simplify finalizing scrub on replica
- We can simply call osr.flush() (with pg lock held) to ensure that prior
writes are visible and scrubbable. This avoi... - 04:29 PM Feature #1649 (Resolved): osd: make replay interval a per-pool setting
- Most pools don't need it. Make it a per-pool thing.
This involves a feature bit and refactor of the pg_pool_t e... - 04:14 PM Revision 29899de5 (ceph): osd: PriorSet: acting/up membership implies still alive
- If the osd is in the acting or up sets, we can assume they are still alive,
even though we don't know that for sure, ... - 03:58 PM Revision a1ddec2a (ceph): Merge remote branch 'gh/master' into wip-prior
- Conflicts:
src/osd/PG.cc - 03:37 PM Messengers Feature #1648 (Duplicate): msgr: choose ip to bind to based on network
- Currently we bind to an explicit address or to any, and learn what address to advertise by looking at our first outgo...
- 03:34 PM Feature #1647 (Resolved): mon: robust bootstrap
- Currently mkfs looks like:
- create initial states on each monitor independently
- start them up and they'll fo... - 03:29 PM Feature #1646 (Resolved): mon: catch up on committed items before attempting to join quorum
- This will prevent a mon that is way behind from dragging down the mon cluster when it comes back online.
- 12:45 PM rgw Bug #1645 (Resolved): rgw bucket suspended broken
- code still looks at the pool auid, which is obviously broken
- 11:07 AM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- I think the simplest solution would be:
- for all operations, set an xattr with the last op_seq to write to that ... - 10:59 AM Bug #1632 (Need More Info): osd: crash in dequeue_op
- need osd logs
- 10:59 AM CephFS Bug #1640 (Need More Info): mds: failed assert(trim_to > trimming_pos)
- need logs with 'debug journaler = 20' and 'debug ms = 1' on the mds for this one
- 10:58 AM Bug #1624 (Need More Info): osd crash in HearbeatMap::_check
- 10:57 AM CephFS Bug #1509 (Need More Info): cfuse sometimes hangs after unmount
- 10:56 AM CephFS Bug #1596 (Need More Info): mds crash during ffsb on kernel client in CInode::is_frozen
- 10:55 AM Bug #1609 (Need More Info): osd: failed assert(info.last_complete == info.last_update)
- 10:55 AM Bug #1598 (Resolved): osd: fix lost objects
- merged, along with the teuthology tests
- 10:52 AM CephFS Bug #1603 (Need More Info): ceph-fuse crash during unmount
- have this one going in a loop to catch it with logs
- 10:51 AM Bug #1530 (Need More Info): osd crash during build_inc_scrub_map
- 10:51 AM Bug #1432 (In Progress): libvirt: fix definition for rbd params/sources/etc
- 10:51 AM Bug #1508 (Need More Info): iozone stuck on kernel rbd mount
- 10:10 AM Cleanup #1644 (Resolved): osd: prior_set refactor
- 09:39 AM rgw Bug #1643 (Rejected): radosgw-admin log show should accept --time
- 09:39 AM rgw Feature #1642 (Resolved): radosgw-admin log show --nonzero-only
- Have another flag for radosgw-admin low show like --nonzero-only that only prints a log entry if it will have a nonze...
- 09:38 AM rgw Feature #1641 (Rejected): radosgw-admn log show --bandwidth-only
- Have a flag for radosgw-admin log show like --bandwidth-only that reduces a log line down to {'bytes_sent':<number>, ...
- 12:20 AM Revision f94a44e6 (ceph): OSDMonitor: reweight towards average utilization
- The existing reweight-by-utilization calculation did not take into
account the current weight of an OSD, and depended...
10/20/2011
- 11:28 PM Revision 409c5717 (ceph): coverage: don't remove ceph tarball
- We want to keep it for examining core files, and we're already
fetching it here, once per suite run. - 10:56 PM Revision 49b6c118 (ceph): osd: PG::PriorSet: make debug_pg arg const
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:51 PM Revision fa66e65c (ceph): osd: PgPriorSet -> PriorSet
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:50 PM Revision 7bc855a8 (ceph): osd: PgPriorSet: rename prior_set_affected -> affected_by_map
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:47 PM Revision 78236e4e (ceph): osd: PgPriorSet: remove obsolete comment
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:46 PM Revision 2a870c14 (ceph): osd: PgPriorSet: move prior_set_affected into PgPriorSet
- This is really where it belongs.
Signed-off-by: Sage Weil <sage@newdream.net> - 10:46 PM Revision c2e66fbd (ceph): osd: PgPriorSet: kill whoami; make PG arg strictly optional
- It is only used for the debug output prefix. Make it so we can leave it
out entirely (e.g. for unit tests).
We don'... - 09:12 PM Revision 47e938c0 (ceph): Merge branch 'stable'
- 09:12 PM Revision 2b3bdea9 (ceph): osd: fix requeue_ops
- The ls argument passed to requeue_ops() is a reference, and one of the
methods we call (say, _handle_op) might want t... - 08:59 PM Revision 3b76f9fc (ceph): perfcounters: remove dout
- We can't use this because we're part of libglobal and there is no
g_ceph_context. And i'm too lazy to use cct.
Sign... - 08:58 PM Revision 863e5b04 (ceph): perfcounters: fix unit test
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 08:48 PM Revision 1d002a1e (ceph): Merge remote branch 'gh/wip-unfound'
- 08:16 PM Revision 28df1e91 (ceph): filestore: measure commit interval, latency, journal full count
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:45 PM Revision d2dbae97 (ceph): osd: clean up perfcounter names
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:43 PM Revision d31e78f6 (ceph): filestore: simplify perfcounter lifecycle
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:43 PM Revision b000e4d4 (ceph): filestore: simplify, clean up perfcounters
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:34 PM Revision d26488bc (ceph): perfcounters: fix addition/removal
- We are not responsible for deleting removed perfcounters.
Add debugging.
Signed-off-by: Sage Weil <sage.weil@dreamh... - 06:33 PM Revision 7207b819 (ceph): filestore: fix perfcounter definition
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:59 PM Revision 53ad579e (ceph): filestore: fix logger start
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:54 PM Revision d3366ccc (ceph): Merge remote-tracking branch 'github/master' into wip-swift-fix
- 05:52 PM Feature #1420 (Resolved): build release rpms
- this is done the extent that i am willing to spend time on it. hopefully suse will show up at some point and improve...
- 05:45 PM Bug #1636 (Resolved): reweight-by-utilization does not choose good weights
- The existing reweight-by-utilization code didn't make sense - commit:f94a44e688883f2db0971435a5333a8b60c77dec fixes t...
- 04:11 AM Bug #1636 (Resolved): reweight-by-utilization does not choose good weights
- there's a problem distributing the data evenly over all devices.
i'm using v0.36 and have a test setup with two host... - 05:22 PM CephFS Bug #1640 (Resolved): mds: failed assert(trim_to > trimming_pos)
- This happened with bonnie++ on cfuse in teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/729/remote/ubuntu...
- 05:09 PM RADOS Feature #1639 (New): osd: guard against bad objects in cls map functions
- Got this when I accidentally set a bad locator:...
- 04:21 PM Revision 288ccc88 (ceph): perfcounters: clean up interface a bit
- No logger_ prefix necessary.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:21 PM Revision daea03ef (ceph): perfcounters: use simple names
- We don't need to uniquely identify ourselves in the global namespace with
the PerfCounter name.. only in the current ... - 02:15 PM Bug #1624: osd crash in HearbeatMap::_check
- commit:2b3bdea9f7bcf9e9f8d4328f62d82ff43e996b3a fixes at least some of these....
- 01:45 PM Bug #1624: osd crash in HearbeatMap::_check
- running this in a loop with logs to try ot catch it
- 12:02 PM Bug #1624: osd crash in HearbeatMap::_check
- And teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/730/remote/ubuntu@sepia50.ceph.dreamhost.com/log/osd....
- 12:00 PM Bug #1624: osd crash in HearbeatMap::_check
- And teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/727/remote/ubuntu@sepia27.ceph.dreamhost.com/log/osd....
- 11:57 AM Bug #1624: osd crash in HearbeatMap::_check
- Happened again in teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/726/remote/ubuntu@sepia41.ceph.dreamhos...
- 02:01 PM Feature #1630 (Resolved): Monitor journal fullness (bytes used, size) via perfcounters
- 01:45 PM Bug #1635 (Duplicate): osd hit suicide timeout in heartbeat_map thread
- 01:28 PM Bug #1588: blogbench on kclient possibly made machine die
- Happened again today - just more transactions timing out in the logs.
- 01:23 PM Bug #1633: osd crash in CryptoKey::decrypt
- Happened again while thrashing in teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/732/remote/ubuntu@sepia...
- 11:52 AM Bug #1530: osd crash during build_inc_scrub_map
- This happened again during cfuse on ffsb (teuthology:~teuthworker/archive/nightly_coverage_2011-10-20/694).
- 09:27 AM Bug #1638 (Won't Fix): Can't create object with large xattrs in a single operation (on extN)
- A single compound operation that does:
- create
- setxattr (small enough to fit but large enough to fill in the e... - 04:45 AM Revision f25879ac (ceph): encoding: add optional features
- Update encode macros to allow a feature bitmask to be passed through
to a classes encode() method.
Signed-off-by: Sa... - 04:22 AM Feature #1637 (Duplicate): OSDs running full take down other OSDs
- this issue has a relation to #1636.
in my test setup of v0.36 when one OSD runs full it gets taken down.
this start... - 04:14 AM Revision 0aa40ea0 (ceph): assert: no 0x before thread id
- There's no 0x prefix in the log lines either. This makes it easier to
copy/paste word and search.
Signed-off-by: Sa... - 03:48 AM Revision 0f0c5947 (ceph): osdmap: uninline big stuff
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:46 AM Revision a71455c8 (ceph): rgw: properly handle cleaning up of listings
- If a listing you get back from the OSD consists only of
non-existent entries, you still need to handle it and resume ... - 12:46 AM Revision 470742d8 (ceph): cls_rgw: move stat update code after error checks in complete_op
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 12:46 AM Revision 6dc0da4c (ceph): cls_rgw: implement a dir_suggest_changes function.
- This takes a bufferlist of suggested changes to the directory, trims
out any sufficiently old tags, and then applies ... - 12:46 AM Revision 952ebbae (ceph): cls_rgw: add constructors to data structs; don't leak tags on races
- We were leaking tags on races before, since we cut out of the function
before clearing the tag. We don't do that any ... - 12:45 AM Revision 9496732d (ceph): rgw: write and use the check_disk_state function
- This is used to check the actual on-disk state, and encode
suggested updates for the index.
Then cls_bucket_list send...
10/19/2011
- 11:47 PM Revision a5ada568 (ceph): rgw: fix bad snprintf
- 10:35 PM Revision 5de847f3 (ceph): .gitignore: add test_filestore_idempotent
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 10:35 PM Revision b57e8967 (ceph): test_filestore_idempotent: initialize var
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 09:58 PM Revision 7f5f1ec1 (ceph): rgw: implement swift metadata POST
- 04:14 PM Revision cf333152 (ceph): Merge branch 'stable'
- Conflicts:
src/mon/OSDMonitor.cc
src/osd/OSD.cc - 03:15 PM Bug #1526: log bound mismatch after thrashing with bonnie
- another occurance, running on swab. This may have led to pg version reset....
- 03:02 PM Bug #1635 (Duplicate): osd hit suicide timeout in heartbeat_map thread
- This was while thrashing with radosbench, during peering, with osds 3 and 6 marked out.
From teuthology:~teuthworker... - 02:46 PM Bug #1634 (Can't reproduce): osd: crash decoding non-existent object_info_t
- From teuthology:~teuthworker/archive/nightly_coverage_2011-10-19/680/remote/ubuntu@sepia28.ceph.dreamhost.com/log/osd...
- 12:05 PM Bug #1633 (Resolved): osd crash in CryptoKey::decrypt
- From teuthology:~teuthworker/archive/nightly_coverage_2011-10-19/682/remote/ubuntu@sepia72.ceph.dreamhost.com/log/osd...
- 11:59 AM Bug #1632 (Can't reproduce): osd: crash in dequeue_op
- During ffsb:...
- 11:26 AM Bug #1631 (Can't reproduce): osd: failed assert(repop_queue.front() == repop)
- This happened on two osds during a multiple_rsync workunit (teuthology:~teuthworker/archive/nightly_coverage_2011-10-...
- 10:51 AM Feature #1630 (Resolved): Monitor journal fullness (bytes used, size) via perfcounters
- 05:33 AM Revision b297d1ed (ceph): osdmap: make encoding based on features
- Instead of relying on the caller to decide whether encode_old_client()
is appropriate, pass in the feature set and en... - 05:26 AM Revision 6e2018ce (ceph): osd: normalize encoding of pg_pool_t
- Instead of using a cumbersom C struct, move members into pg_pool_t and
use normal encode/decode methods.
Signed-off-... - 05:26 AM Revision cee1b27f (ceph): crush: clean up encoder/decoder
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:25 AM Revision 954d3f9b (ceph): use WRITE_CLASS_ENCOER macro when possible
- 05:25 AM Revision 9d93bfce (ceph): encoding: WRITE_CLASS_ENCODER_MEMBER -> WRITE_CLASS_MEMBER_ENCODER
- Signed-off-by: Sage Weil <sage@newdream.net>
10/18/2011
- 11:52 PM Revision 83914d22 (ceph): test_filestore_idempotent: simple tool to generate a worklaod of non-id...
- Generate a workload of operations that are non-idempotent. These are:
transaction {
clone A -> A.($n-1)
writ... - 11:52 PM Revision 3e92aace (ceph): filestore: tolerate EEXIST on mkcoll when not-btrfs
- For non-btrfs file systems we should tolerate EEXIST because we may
replay the event more than once.
Signed-off-by: ... - 11:41 PM Revision ba165fec (ceph): rgw: some swift obj metadata related fixes
- 11:18 PM Revision 13b0bbb3 (ceph): mds: handle xattrs on inode creation
- Allow mknod, mkdir, symlink, create to provide xattrs for the new
inode. This will be used by the kclient to set ACL... - 11:04 PM Revision 7ea07832 (ceph): radosgw-admin: fix conflict with KeyType in libnss
- rgw/rgw_admin.cc:459:6: error: using typedef-name 'KeyType' after 'enum'
/usr/include/nss3/keythi.h:69:3: error: 'Key... - 08:28 PM Revision ed5e4341 (ceph): rgw: add content-type to index dirent
- 06:42 PM Revision da6cdfdd (ceph): osd: PgPriorSet: cur -> probe
- Rename cur to probe, the set of OSDs we need to probe in order to
successfully peer.
Signed-off-by: Sage Weil <sage@... - 06:40 PM Revision 4e5242e0 (ceph): osd: PgPriorSet: restructure lost checks for prior set
- When we add down osds to the cur set, we block peering because there
are OSDs that may have data we need and they are... - 06:01 PM Revision 298dbbe6 (ceph): rgw: workqueue suicide timeout is infinity
- 04:49 PM Bug #1629 (Can't reproduce): pgs stuck degraded (only mapped to 1 osd)
- From teuthology:~teuthworker/archive/nightly_coverage_2011-10-18/636/teuthology.log:...
- 04:42 PM Bug #1624: osd crash in HearbeatMap::_check
- argh, the tarball is already gone:
# wget http://ceph.newdream.net/gitbuilder/output/sha1/e6dbd7141bd8b4403f3b931f... - 12:00 PM Bug #1624 (Resolved): osd crash in HearbeatMap::_check
- Logs with debugging are in vit:~joshd/thrash_stuck_active4. This happened on osds 0 and 4:...
- 03:35 PM Bug #1628: segfault attempting to map an rbd snapshot
- This is a bug in the rbd command line tool - it accepts snapname but doesn't use it for map/unmap. Additionally, it d...
- 03:18 PM Bug #1628 (Resolved): segfault attempting to map an rbd snapshot
- ...
- 03:08 PM Bug #1626: ceph-mon HA not working right; all must be up
- Sorry to dribble this in: it seems with one mon down and voted out, "ceph -s" takes <1sec 66% of the time, ~3sec 33% ...
- 03:07 PM Bug #1626: ceph-mon HA not working right; all must be up
- Oh sorry, what I see with vstart is a 10-second timeout until the mons vote mon.c out. This is *not* what Carl report...
- 03:05 PM Bug #1626: ceph-mon HA not working right; all must be up
- Carl saw it originally. Easy to repro with vstart:...
- 02:48 PM Bug #1626: ceph-mon HA not working right; all must be up
- where did you see this?
- 02:28 PM Bug #1626 (Can't reproduce): ceph-mon HA not working right; all must be up
- If mon.gamma is down, "ceph -s" hangs trying to connect to all three ceph-mon. The paxos majority rule system does no...
- 02:52 PM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- FWIW even if we know what not to replay, we could still be screwed with ext4 (which does not commit everything in ord...
- 02:25 PM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- Tommi Virtanen wrote:
> Isn't the idempotency in that case "clone foo_head -> foo_2 IFF foo_2 does not exist" ?
T... - 02:06 PM Bug #213: non-idempotent transactions (clone) under ext3 may not replay correct result
- Isn't the idempotency in that case "clone foo_head -> foo_2 IFF foo_2 does not exist" ?
- 02:37 PM Bug #1627 (Can't reproduce): ceph-mon memleak if ceph-osd cluster ip is not reachable, but public...
- ...
- 02:25 PM Feature #1625 (Rejected): changing ceph-mon ip address needs monmap change on every mon machine
- Moving mon.{alpha,beta,gamma} to new IP addresses was a fairly convoluted process. This would be nice if it was simpl...
- 02:15 PM Feature #641: allow logs to be piped to an external program
- This is feature creep. If you want to process the logs asap in another process, just have it get ceph stdout as stdin...
- 11:19 AM Bug #1620 (Resolved): rgw suicide due to heartbeat timeout
- Fixed, commit:298dbbe64f8b0738ec58db43782813d0686717c7. Basically a 0 value for the rgw suicide timeout should do the...
- 11:01 AM Bug #1588: blogbench on kclient possibly made machine die
- This happened again yesterday and today with different machines. Both times, the only unusual thing in kern.log was t...
- 01:59 AM Revision 0a027599 (ceph): osd: PgPriorSet: simplify (and change) CRASHED logic
- Any single OSD from a given interval surviving is sufficient to ensure
that an ACKed write during that interval was c... - 01:57 AM Revision f7ef94d3 (ceph): osd: PgPriorSet: update comment terms a bit
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:51 AM Revision bbb06d34 (ceph): osd: do not short-cut up_thru update for new PGs
- Commit e731885d2550ee985bf875ab5bb5faf28f1693eb made it possible for
a new PG to go active without forcing the OSDs u... - 12:44 AM Revision 57e0ab74 (ceph): osd: PgPriorSet: clean up per-interval var names
- We don't actually use any_lost_now, but it makes the logic easier
to understand to have it there.
Signed-off-by: Sag... - 12:44 AM Revision 53381364 (ceph): osd: PgPriorSet: clean up comments a bit
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:44 AM Revision 33b33f7e (ceph): osd: PgPriorSet: remove unused PG member
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:44 AM Revision 113b7833 (ceph): osd: PgPriorSet: revert start_since_joining check
- Commit 5b78f5db8c200edcc949033e1badae70fecd2e08 added a check to
prevent some sort of badness when osds were marked l... - 12:44 AM Revision 3dda4465 (ceph): osd: PgPriorSet: remove up_thru crap
- This was added way back in 1cf9bebc8e5063f5f311d33e7735bcc9286e98ce,
but as far as I can tell it didn't make any sens... - 12:44 AM Revision f89f4d9b (ceph): osd: PgPriorSet: do not include UP osds in prior.cur
- The up osds are not (directly) relevant since they are not necessarily
members of the PG. We only care about acting ... - 12:09 AM Revision 9dfa1105 (ceph): rgw: fix swift account and containers listing limits
10/17/2011
- 11:48 PM Revision c5638b70 (ceph): osd: PgPriorSet: any_survived -> any_is_alive_now
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:13 PM Revision e6dbd714 (ceph): doc: Change diagram to have radosgw closer to direct rados access.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 10:35 PM Revision 3c90c0d4 (ceph): add singleton lost-unfound
- 10:32 PM Revision 4ec37b23 (ceph): add lost_unfound task
- Also some misc useful bits to ceph_manager.
- 10:21 PM Revision edcd4d97 (ceph): rgw: some more swift fixes
- 09:45 PM Revision 83cf3fef (ceph): Expect 'wrongly marked me down' messages during thrashing
- 09:42 PM Revision bcded7f1 (ceph): ceph: add whitelist for cluster log errors
- Some messages are expected when thrashing osds or creating unfound
objects.
Fixes: #1622 - 09:13 PM Revision 0bad37e3 (ceph): streamtest: do mkfs
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:12 PM Revision 525a610f (ceph): streamtest: print to stdout
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:49 PM Revision 9c956049 (ceph): mkcephfs: copy ceph.conf to /etc/ceph/ceph.conf (when -a)
- You can disable this with --no-copy-conf.
Signed-off-by: Sage Weil <sage@newdream.net> - 05:40 PM Revision fba220ec (ceph): nuke: reset syslog configuration after rebooting
- Previously we removed a file and rebooted without syncing, so the file
was never deleted. - 04:56 PM Bug #1623 (Can't reproduce): ceph-osd fails to bind socket
- ...
- 03:51 PM Revision 9baf5ef4 (ceph): ceph.spec: don't chkconfig
- This was fighting with suse insserv. Still needs some cleanup.
Signed-off-by: Sage Weil <sage@newdream.net> - 03:50 PM Revision 21d941e8 (ceph): ceph.spec: work around build.opensuse.org
- The redhat-rpm-config isn't installed on build.opensuse.org, which means
the processor is set to i386 instead of some... - 03:49 PM Revision 195a484b (ceph): ceph.spec: capitalize first letter to make rpmlint happy
- Signed-off-by: Sage Weil <sage@newdream.net>
- 03:35 PM Revision a6f3bbb7 (ceph): v0.37
- 03:27 PM Bug #1473 (Resolved): osd assert failure: FAILED assert(0 == "oi disagrees with stat, or error c...
- 03:26 PM Bug #1473: osd assert failure: FAILED assert(0 == "oi disagrees with stat, or error code on stat")
- At least the recent instances of this were probably caused by the btrfs xattr bug (#1612).
- 03:25 PM Bug #1486 (Resolved): osd: 0-length meta/pginfo_* files
- 10:10 AM Bug #1486: osd: 0-length meta/pginfo_* files
- I saw this on alexandria, and it was caused by:
1- EMFILE (too many open files)
2- filestore wasn't assering on... - 03:20 PM Bug #1612 (Resolved): osd/PG.cc: 3839: FAILED assert(missing[oid].need <= v)
- This was caused by a btrfs xattr bug. I got a patch back from josef and pushed it to the dho kernel.
- 03:05 PM Feature #1622 (Resolved): teuthology: whitelist ceph.log entries
- Implemented in teuthology and whitelisted 'wrongly marked me down' messages for thrashing jobs in the suite.
- 10:08 AM Feature #1622 (Resolved): teuthology: whitelist ceph.log entries
- Need to be able to do this to make certain tests pass. notably thrashing and the new lost_unfound.
- 03:37 AM Revision ca8f6036 (ceph): osd: fix assemble_backlog
- This was written assuming that le->prior_version wouldn't be the version
that we have locally on disk. Not always tr... - 03:37 AM Revision 2fdec7b8 (ceph): osd: fix add_next_event Missing::item::have
- The missing set should be accurate up to the current point in the log. The
log_tail has no bearing on that, nor does...
Also available in: Atom