Activity
From 03/25/2011 to 04/23/2011
04/23/2011
- 11:02 PM Revision 52b74811 (ceph): mon: fix up osd health report
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:08 PM Revision a9d12cb5 (ceph): mon: fix up pg health report
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:07 PM Revision 70640bf1 (ceph): mon: generalize health check a bit
- Any PaxosService can return a status (or not).
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 09:32 PM Feature #975: osd: use fallocate(FL_PUNCH_HOLE) for zero
- This doesn't appear to be in btrfs yet anyway.. let's wait!
- 09:27 PM Bug #1003 (Resolved): mon: health should show % degraded, unfound (and warn appropriately)
- 08:09 PM Revision fcf095c7 (ceph): Merge branch 'master' of ssh://ceph.newdream.net/home/sage/ceph.newdrea...
- 01:38 PM Feature #958 (Resolved): rbd: map/unmap kernel rbd device via rbd(?) tool
- 12:22 AM Revision 4368e974 (ceph): Makefile.am: add -fno-strict-aliasing
- Until we audit every place we have a typecast to make sure it's safe,
this should be off.
Signed-off-by: Colin McCab... - 12:17 AM Revision 4fe53fca (ceph): Makefile.am: Wpointer-arith, Wstrict-null-sentinel
- Add Wpointer-arith, Wstrict-null-sentinel, Winit-self
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
04/22/2011
- 11:52 PM Revision 793034c6 (ceph): v0.27
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:41 PM Revision 4b547cbb (ceph): Makefile: add some new warnings to CXXFLAGS
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:39 PM Revision 268f1897 (ceph): clitests: fix osdmap unit test
- Broken by removal of casdata in 8cd3f1cd2546a61bb5402fa374de7f37147dc3a0.
Signed-off-by: Sage Weil <sage.weil@dreamh... - 11:20 PM Revision 474be659 (ceph): mds: don't daemonize when doing journal reset.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 10:31 PM Revision 625e7851 (ceph): rbd: rename kernel add/remove to map/unmap
- Unmap takes a device now, and uses the major number
to find the right sequence number in sysfs.
Also updated the man... - 08:39 PM Revision f5e20331 (ceph): Merge remote branch 'origin/next'
- 08:34 PM Revision 4cc88f62 (ceph): common, mount.ceph: move functions for working with secrets into secret.h
- Use sprintf instead of safe_cat, since we're just writing a string once.
Signed-off-by: Josh Durgin <josh.durgin@dre... - 08:34 PM Revision 20719c7a (ceph): rbd: use rados' init_with_config because we already called common_init
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 08:34 PM Revision bffb9edf (ceph): addr_parsing: move implementation out of header
- Also include missing headers and make it linkable with C,
since it's duplicated from mount.ceph.
Signed-off-by: Josh... - 08:34 PM Revision def36668 (ceph): addr_parsing: rename mount_resolve_dest to resolve_addrs and don't modi...
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 08:34 PM Revision 2ff3c82b (ceph): addr_parsing: use strtok_r to be safe
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 08:34 PM Revision 98698596 (ceph): mount.ceph: use common implementations of secret handling and address r...
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 08:34 PM Revision c87fc696 (ceph): rbd: add commands for manipulating the kernel sysfs interface
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 07:30 PM Revision 88ebb304 (ceph): Merge remote branch 'origin/next'
- 06:37 PM Revision 8cd3f1cd (ceph): osd: drop 'casdata' bucket by default
- We don't use it for anything. It just means we create a bunch of useless
PGs.
Signed-off-by: Sage Weil <sage@newdre... - 06:37 PM Revision cb8237e1 (ceph): testlibrbd: fix signed/unsigned comparisons
- testlibrbd.c: In function 'write_test_data':
testlibrbd.c:191: warning: comparison between signed and unsigned intege... - 05:33 PM Revision b419f3d7 (ceph): filestore: lfn_unlink removes the entry as last step
- so that if it crashes in the middle of operation we're not left
with a hole in the list of objects. In case of a cras... - 01:37 PM CephFS Bug #1023 (Resolved): just-recovered mds fails journaler assert (to > trimming_pos);
- 3 mdses configured for standby-replay. mds.1 was active, the other two?!? (not sure) were apparently following it. ...
- 12:17 PM Linux kernel client Bug #970: Kernel crash (cause?: lots of small files)
- probably a memory leak?
- 12:13 PM CephFS Bug #1022: every mds crash: Program terminated with signal 11, Segmentation fault.
- This looks like damage from a bug in the session journaling. Can you dump a copy of your journal so we can take a cl...
- 12:34 AM Revision 00ffafa8 (ceph): cmds: make journal-check mode run in the foreground.
- daemonize does a bunch of stuff we don't want to do.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 12:10 AM Revision 804a6d6c (ceph): mds: massage handling of oneshot-replay.
- We accidentally broke it in our last round of changes to support
standby-replay; patch it up again.
Signed-off-by: G...
04/21/2011
- 08:18 PM Revision 7979fa5c (ceph): OSDMon: make the defaults in reweight-by-utilization work.
- an oload <= 100 fails out; set the new default to 120.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 05:06 PM CephFS Bug #1006 (Resolved): clean up cmds journal-check mode
- Well, it will still take over the mds.a symlink if you start it up as a pre-existing MDS, but you shouldn't do that. ...
- 06:22 AM CephFS Bug #1022 (Can't reproduce): every mds crash: Program terminated with signal 11, Segmentation fault.
- A few seconds after startup, all my MDSes crash with the following message:...
- 03:50 AM Revision 8109f745 (ceph): testlibrbd: fix signed/unsigned comparisons
- testlibrbd.c: In function 'write_test_data':
testlibrbd.c:191: warning: comparison between signed and unsigned intege... - 03:50 AM Revision c8059372 (ceph): osd: drop 'casdata' bucket by default
- We don't use it for anything. It just means we create a bunch of useless
PGs.
Signed-off-by: Sage Weil <sage@newdre... - 12:11 AM Revision 1cff807c (ceph): filestore: lfn fixes
- also bumping up max object len to 2048, we'll probably want to change
that to a higher number (4096?), need to make s...
04/20/2011
- 10:16 PM Revision 4e21db6e (ceph): auth: const cleanup
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:35 PM Revision 1421d70d (ceph): obysnc: use python2.5 compatible exception syntax
- Apparently using "as" in "except" blocks is a new (and not
backwards-compatible) thing in Python 2.6.x. Who knew?
Si... - 08:54 PM Revision 3d97c33b (ceph): PGMon: use the config observer framework
- We now register observers on mon_osd_full_ratio and nearfull_ratio
to update the OSD full/nearfull lists.
Signed-off... - 08:51 PM Revision 19408759 (ceph): filestore: fix getxattr
- 08:51 PM Revision f940ee86 (ceph): ProfLogger: add NULL to KEYS in the config observer.
- I believe that without the null pointer it will overrun into
uninitialized memory.
Signed-off-by: Greg Farnum <grego... - 08:51 PM Revision 0e828416 (ceph): config: add comment clarifying virtual destructor.
- Just in case anybody else is like me and gets confused.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 08:49 PM Revision e209db4a (ceph): PGMonitor: unregister the config obs in ~PGMonitor
- Using the new unregister call, unregister the config observer in
~PGMonitor.
Signed-off-by: Colin McCabe <colin.mcca... - 08:38 PM Revision 527ecd4d (ceph): config: add remove_observer
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:58 PM Revision 5b218b27 (ceph): filestore: introduce long-file-names
- still needs testing, adjusting of constants and using a real hash function
- 05:58 PM Revision 7dc5a7f9 (ceph): PGMap: fill in full OSD sets on decode.
- Previously these sets were only modified when applying an incremental,
which might have caused problems when restarti... - 01:26 PM Feature #1010 (Resolved): mon: adjust mon-osd-[near]full-ratio on the fly
- This seems to be working in master on commit:3d97c33bea5144cbe5ed1b40704f74fff1ec0fb1.
- 10:19 AM Subtask #843 (Resolved): Decouple dout from g_conf
- implemented by commit:42a7013907e312d864cfcf591cca932cf9f343b2
- 10:19 AM Bug #1011 (Resolved): config observer structure to eliminate SIGHUP requirement
- implemented by commit:6e49c06ca90dc7aa696cbe298b4362dd4cb124ba, commit:8ebbd60a865605e7cc88e07b2b3f2f84521934a9
- 09:25 AM Bug #1021 (Can't reproduce): osd: _process_pg_info FAILED assert(pg->log.tail <= pg->info.last_co...
- ...
- 07:34 AM CephFS Bug #1019 (Resolved): mds: ~MDSlaveUpdate while on LogSegment list
- 04:55 AM Revision 07f287c6 (ceph): Merge remote branch 'origin/classdist'
- 12:59 AM Revision 8ebbd60a (ceph): conf: reopen proflogger on conf change
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:28 AM Revision 6e49c06c (ceph): conf: add observer framework
- Objects can now register as configuration observers interested in a
subset of the configuration keys. The observers w...
04/19/2011
- 11:11 PM Linux kernel client Feature #42 (Resolved): Resize of rbd image
- commit:e56433a0330b61e56489779e21f68fdc4c45f5be
- 10:33 PM Revision 36f00685 (ceph): cauthtool: -C not -c in man page
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:23 PM Linux kernel client Feature #819: robust nfs re-export
- There are patches queued for .40 that also try LOOKUPINO on unconnected fh's. This all needs testing.
- 10:22 PM Linux kernel client Bug #811 (Can't reproduce): rbd warning and crash
- 10:21 PM Cleanup #892 (Resolved): osd: remove class distribution infrastructure?
- 10:19 PM Tasks #584 (New): do throughput scaling tests on sepia
- 10:03 PM Revision 9a38acf0 (ceph): dout: remove DOUTSB_FLAG_STDOUT
- This was used to implement log_to_stdout, but that doesn't exist any
more. Instead of we have log_to_stderr and the o... - 10:03 PM Revision 433deec0 (ceph): dout: make dout a member of md_config_t
- Since logging options are per-config, logically DoutStreambuf instances
should also be per-config. This also allows u... - 10:02 PM Revision 0b6a023f (ceph): PGMap: add redo_full_sets() function for when ratios change.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 09:55 PM Revision af92e8bf (ceph): PGMap: always decode pg_remove.
- pg_remove has been included for longer than we've had versions
in the struct, so this check for end is useless -- if ... - 09:55 PM Revision a2613f82 (ceph): PGMap: add [near]full_ratio to the Incremental and encoding.
- This way these ratios can be distributed by map updates.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 09:55 PM Revision 515c7677 (ceph): osdmon: Clean up debug output on map full flag.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 09:32 PM Revision f9056d0d (ceph): osd: better debug output on replay completion
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:13 PM Revision 634dfc90 (ceph): mkcephfs: allow a prebuild osdmap to be specified
- Otherwise we'll create one with osdmaptool --createsimple with the default
generic settins.
Signed-off-by: Sage Weil... - 08:57 PM Revision ad71955a (ceph): cfuse, mds, dout: Do not create "rank" symlinks for log files.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 07:05 PM Revision 4428d1ec (ceph): Revert "Revert "autoconf: Complain if tcmalloc is not found.""
- This reverts commit 05c281bfa9e9d69ea3d0197590950c8e6845a13a.
This should be okay now. - 07:05 PM Revision f6179fc3 (ceph): debian: Handle missing tcmalloc on Debian lenny.
- lenny doesn't have a suitable libgoogle-perftools-dev, and
release.sh edits it out of build-deps. Detect that and tel... - 07:05 PM Revision 0d98a62c (ceph): debian: Build without tcmalloc on non-i386/amd64.
- This is not strictly needed as of 05c281bfa9e9d69ea3d0197590950c8e6845a13a,
but that reverting is hopefully only temp... - 06:59 PM Revision bd7922a1 (ceph): remove old debian build helper scripts.
- This is now handled by ceph-build.git
Signed-off-by: Sage Weil <sage@newdream.net> - 06:33 PM Revision 7c04f81c (ceph): mon: remove class distribution infrastructure
- This is now the admin's job. Removes a lot of code with limited testing
and coverage.
We rev the internal monitor p... - 06:11 PM Revision deb0ee90 (ceph): osd: move ClassHandler.{cc,h} to osd/
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:09 PM Revision 964a0a6e (ceph): osd: load classes from a fixed location off disk
- Simplify the class loading code to just load out of a local directory.
Do not request classes from the monitor.
Sign... - 05:28 PM Revision 54284c0a (ceph): common: add environment.cc/environment.h
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:51 PM Bug #1020 (Duplicate): osd: replay_queued_ops crash
- duplicates 990. Working on a peering/recovery refactor which should take care of this.
- 02:35 PM Bug #1020 (Duplicate): osd: replay_queued_ops crash
- ...
- 04:50 PM Bug #906: clustered mds: lchown not setting uid/gid
- Here's an idea: run the autotest say 10 times (after the test, ssh to the sepia machines and ensure they've rebooted,...
- 04:43 PM Bug #906: clustered mds: lchown not setting uid/gid
- Still unable to reproduce this locally, and running it again on the autotest cluster it didn't fail.
It's possible... - 04:37 PM CephFS Bug #1006 (In Progress): clean up cmds journal-check mode
- Okay, it's more broken than that -- somehow it ends up in standby-replay instead of oneshot-replay!
- 04:25 PM Revision d55399ff (ceph): mds: remove MDSlaveUpdate from list on deletion
- These are added to the LogSegment list on the slaves, but also need to be
removed from that list when we replay a COM... - 11:43 AM Linux kernel client Bug #1014: kclient can't complete fsstress on multi-mds cluster
- Not high priority right now, so unassigning.
- 09:13 AM CephFS Bug #1019: mds: ~MDSlaveUpdate while on LogSegment list
- commit:d55399ffec224206ea324e83bb8ead1e9ca1eddc
- 08:52 AM CephFS Bug #1019 (Resolved): mds: ~MDSlaveUpdate while on LogSegment list
- ...
- 08:54 AM Bug #1015 (Rejected): ceph 0.26, mkcephfs : ERROR: error creating empty object store
- You need to create the directory yourself and this is deliberate. It prevents administrators from not fully setting u...
- 08:31 AM Bug #1018 (Can't reproduce): error on building ceph on red hat 5.5
- building ceph on rh 5.5 i have obtained this error.
if g++ -DHAVE_CONFIG_H -I. -I. -I. -Wall -D__CEPH__ -D_FILE... - 12:54 AM Revision 87def518 (ceph): common_init: move version printout to common_init
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:32 AM Revision ae0f5f03 (ceph): interval_set: slightly better API for erase
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:18 AM RADOS Bug #1017 (Closed): ceph 0.26 ,mkcephfs --crushmap crush.new ,wait for very long time,mds stat i...
- Hi ,
ceph 0.26 + btrfs +ubuntu 10.04 x86_64;
we want to use a special crushmap when do mkcephfs,but wait for very v... - 12:12 AM RADOS Bug #1016 (Closed): ceph 0.26,crushmap change,mount fail.
- Hi ,
ceph 0.26 ,change crushmap,then at a client ,mount ceph ,but fail.ceph.conf/crush.new.txt/crush.origin.txt/ a...
04/18/2011
- 11:51 PM Bug #1015 (Rejected): ceph 0.26, mkcephfs : ERROR: error creating empty object store
- Hi ,
when do mkcephfs ,if it doesn't create osdN directory manually, then ,mkcephfs fail.
1. ceph 0.26 + btrfs +ubu... - 11:26 PM Revision faffcc56 (ceph): Change some global config options to env vars
- Options that are inherently global, like malloc settings, and also
inherently debugging or profiling settings should ... - 11:26 PM Revision e4e2b742 (ceph): Merge commit '8038c491ba90a8cbcd569e84d4cafc8bbdff81d5' into next
- 11:23 PM Revision c93c6619 (ceph): Merge remote branch 'origin/stable' into next
- 10:32 PM Revision 2e41f084 (ceph): journaler: fix do_flush adjustment
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:30 PM Revision bb4cfabc (ceph): osd: assert on query on deleting pg
- This shouldn't happen; assert if it does.
Signed-off-by: Sage Weil <sage@newdream.net> - 10:29 PM Revision 67721b00 (ceph): osd: do not create pgs on query
- We do not need to create the pg on query. If we are a new replica we can
create it when we get the PGLog activate me... - 10:06 PM Revision b05f1ba4 (ceph): journaler: truncate/zero ahead of write position
- Remove/zero objects N periods ahead of the journal write position. This
ensures that when we reprobe the journal len... - 10:06 PM Revision 974d859e (ceph): filer: remove whole objects on zero()
- Make filer::zero() remove any whole objects. This is required by the
Journaler, given the way it probes the journal ... - 10:06 PM Revision 0f0c1c85 (ceph): interval_set: implement erase(iterator)
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:06 PM Revision eaf52413 (ceph): Merge branch 'next'
- 09:34 PM Revision e6393525 (ceph): journaler: check return code on _finish_write_head.
- If we get an error code and assume we successfully wrote the head,
there are going to be all kinds of issues on repla... - 08:55 PM Revision 68863bb4 (ceph): osd: make ZERO on non-existent object a no-op
- Fixes bug where oi.size gets out of sync with the object size because we
actually write zeros. (This explains #933.)... - 08:12 PM Revision ab9dc122 (ceph): vstart tests: specify ceph.conf
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:12 PM Revision 42a70139 (ceph): dout: remove references to g_conf
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:54 PM Revision 878f4a45 (ceph): rgw: return appropriate error for unresolvable email in acl
- 06:53 PM Revision b8e791bc (ceph): config: reorder config_option struct members
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:52 PM Linux kernel client Bug #1014 (Closed): kclient can't complete fsstress on multi-mds cluster
- cfuse is doing great with fsstress on multi-mds systems now, but the kclient won't finish. Analysis is needed.
- 06:48 PM Bug #910 (Resolved): Multi-MDS Ceph does not pass fsstress
- I've been unable to break this under cfuse in master all day.
- 06:47 PM CephFS Tasks #1002 (Resolved): Assert failure in Locker::handle_file_lock
- 11:36 AM CephFS Tasks #1002: Assert failure in Locker::handle_file_lock
- I suspect this got fixed up by some of the other changes, but want to leave it open for a little longer just to make ...
- 06:44 PM Revision 3f275bcf (ceph): clitests: eliminate use of old-style section name
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:44 PM Revision 756a2acc (ceph): config: remove unused OPT_NONE type
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:44 PM Revision 8038c491 (ceph): clitests: fix radosgw_admin test
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:41 PM Revision 6058a36c (ceph): MDS: move slave rename xlock handling before finish_export_inode.
- finish_export_inode changes states! That's not good for our checks,
so just handle unpinning and stuff before we fini... - 05:41 PM Revision 14dd2994 (ceph): improve debug printing
- 05:41 PM Revision d8579833 (ceph): mds: Unify migration-handling code in _commit_slave_rename.
- We need to handle locks and pins on exported inodes but we
were using a separate if block with its own (non-matching!... - 05:40 PM Revision 6bd20815 (ceph): mds: _commit_slave_rename needs to drop auth_pins for exported xlocks.
- Otherwise these pins are never dropped from the inode since we
don't go through our normal xlock teardown code. Now w... - 05:40 PM Revision 1a6f4376 (ceph): MDS: Make _rename_apply inode import auth_pinning more intelligent.
- We don't want auth_pins on the locallocks (they're never auth_pinned)
and we only want new auth_pins that are for loc... - 05:40 PM Revision 478c6173 (ceph): mds: If we're a slave, clean up xlocks when we export an inode.
- Because we can do an inode import during a rename that skips the usual
channels, we were getting into an odd state wi... - 05:40 PM Revision 5299aabe (ceph): mds: properly drop imported xlocks.
- Because we can do an inode import during a rename that skips the usual
channels, we were getting into an odd state wi... - 05:36 PM Revision 97e357c4 (ceph): MDS: Server takes auth_pins for xlocks on imported inodes.
- Should fix #934.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 05:15 PM Revision 216fd776 (ceph): objecter: resub ops on full->nonfull transition
- This was broken a while ago during the last refactor. Whoops! Clean it
up to be smarter (and work at all).
Signed-... - 04:58 PM Revision 879adb61 (ceph): Merge remote branch 'origin/stable'
- Conflicts:
src/osdc/Journaler.cc - 04:57 PM Revision c966410f (ceph): osd: show "full" or "nearfull" in osdmap summary line
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:56 PM Revision fa7061d2 (ceph): Merge branch 'rgw_uid'
- 03:56 PM Revision 796528c3 (ceph): rgw: remove get_user_info() and clean up
- rename all the get_uid_by_* to get_user_info_by_*, remove get_user_info()
and call the appropriate function instead (... - 03:32 PM Revision d8fe208d (ceph): rgw: store user info on all indexes in the same format
- this breaks backward compatibility, we'll have to deal with that
later. - 03:15 PM Revision 11f1e2ef (ceph): rgw_admin: can lookup user by access key
- 02:54 PM CephFS Bug #873 (Resolved): MDS journaler needs to truncate ahead of writes
- commit:b05f1ba4c46976bd96a1df8a582dffd89860eb34
- 10:53 AM CephFS Bug #873 (New): MDS journaler needs to truncate ahead of writes
- 02:03 PM Bug #1013 (Can't reproduce): filejournal buffer::rebuild crash
- ...
- 12:41 PM rgw Bug #982 (Resolved): radosgw should fail with UnresolvableGrantByEmailAddress on add_email_grant ...
- Fixed, commit:878f4a45ceebd3038a6983a97064b0fdcfec11fe.
- 10:12 AM rgw Bug #982: radosgw should fail with UnresolvableGrantByEmailAddress on add_email_grant to unknown ...
- At this point, instead of returning an error, we just ignore the bad entry.
- 11:35 AM Tasks #1005 (Resolved): xlock is not unpinning during rename across MDSes
- Looks like this is fixed.
- 11:20 AM Bug #1012 (Rejected): Autotest: Measure RADOS IO performance under read and write loads
- We've had sporadic but steady reports about writes starving reads. Write an autotest to measure various kinds of RADO...
- 11:00 AM Feature #974 (Resolved): autotest: include ffsb benchmarks
- 10:56 AM Bug #935 (Can't reproduce): ceph client heavy loading in Metadata server
- 10:47 AM Bug #1011 (Resolved): config observer structure to eliminate SIGHUP requirement
- 10:11 AM rgw Bug #980 (Resolved): radosgw returns access_key instead of user_id when fetching acl
- Fixed as of commit:796528c3db66d63d0bd75c880cb871d5e4f0322a.
- 09:58 AM Feature #1010 (Resolved): mon: adjust mon-osd-[near]full-ratio on the fly
- The nearfull and full ratios are set in the PGMap constructor, and the lists for maintained privatley by PGMap. That...
- 09:42 AM Tasks #419 (Closed): 24/7 testing
- 09:41 AM CephFS Bug #523 (Can't reproduce): cfuse locks don't wake on mds reconnect?
- 09:37 AM Linux kernel client Bug #1004 (Resolved): hanging writes, bisected to i_state locking change
- commit:283a85d
- 04:58 AM Revision d7789218 (ceph): mount.ceph: behave when CONFIG_KEYS is not compiled in
- In that case we get ENOSYS. This also implies an old version of the client
and that we should fall back.
Signed-off... - 12:42 AM Revision d21bdd6e (ceph): radosgw_admin: Update manpage to new syntax
- Signed-off-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Colin McCabe <cmccabe@alumni.cmu.edu>
04/17/2011
- 04:04 AM Revision 1eccc019 (ceph): MDS: Fix Locker::handle_reqrdlock for xlocked locks.
- We previously dropped the request but that was inappropriate for that
one case because the replica has no way to trig... - 04:03 AM Revision 79cac5ee (ceph): mds: Always _open_parents when opening a new snaprealm
- Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 04:03 AM Revision a028c895 (ceph): mds: don't run all of try_subtree_merge on a rename across MDSes.
- Previously we'd try and do the whole thing, which meant that
the replica got a lock twiddle before it had finished th... - 04:02 AM Revision 6250e82c (ceph): mds: adjust LocalLock can_xlock_local().
- I don't remember why we needed can_xlock_local() to begin with, but
I can tell that adding this get_xlock_by() check ... - 04:02 AM Revision 5a65a04a (ceph): mds: Extend use of find_ino_peers.
- Missed a few places that need it.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Signed-off-by: Sage Weil... - 04:02 AM Revision bea966af (ceph): mds: Make use of find_ino_peers
- Previously we just had to give up on ESTALE. Now
we can attempt to recover!
Signed-off-by: Greg Farnum <gregory.farn... - 04:01 AM Revision 22e8519d (ceph): random commenting
- 04:01 AM Revision ace54db0 (ceph): MDS: Remove inappropriate assert from _logged_slave_rename.
- The slave also can hold some auth pins from locks which the
master has asked it to grab. It's possible we can intelli... - 04:01 AM Revision ac045dc3 (ceph): MDS: Server::handle_slave_rename_prep now accounts for dir snaplock.
- Previously it ignored the auth pin required to hold snap xlock, which
is currently always held for a rename on a dir.... - 04:00 AM Revision 597e30ed (ceph): MDS: Don't move inode to snaprealms if not primary inode.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:59 AM Revision 08bd2ef1 (ceph): MDCache: update assert to account for being a slave.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:59 AM Revision 569cce39 (ceph): Server: push_projected_linkage in _link_remote
- _link_remote_finish will pop the linkage if inc==true, so we'd
better push it to match!
Signed-off-by: Greg Farnum <... - 03:59 AM Revision 5b825c3a (ceph): Server: ensure slave mdses have full dest tree
- We were already taking rdlocks on the source tree, to make
sure that each slave MDS could traverse to the source dent...
04/16/2011
- 10:13 PM Cleanup #1009 (Resolved): osd: push pg message handling code into ReplicatedPG
- There is a lot of code in OSD.cc that does PG-specific work, mostly in the handle_pg_* message handlers. Those in tu...
- 10:07 PM Feature #1008 (Resolved): filestore: handle long object names
- extN and btrfs both limit filenames to 255 characters, but we want to store object names that are much longer than th...
- 10:03 PM Bug #906: clustered mds: lchown not setting uid/gid
- This isn't popping up with single mds... probably a clustering thing.
- 09:42 PM Feature #1007 (Resolved): qa: osd failure and cluster recovery test(s)
- We need tests of OSD failures that verify the cluster is able to recover. Eventually this will need to be fleshed ou...
- 09:33 PM CephFS Bug #994 (Resolved): EOpen reply on non-auth MDS is busted
- I was wrong about the diagnosis before. Pretty sure commit:777bcba0 fixes this.
As for the non-auth caps, I forgo... - 12:20 AM Revision 544ce94a (ceph): rgw: basic support for separate uid and access key
04/15/2011
- 11:32 PM Revision 777bcba0 (ceph): mds: don't skip inodes in journal that may be trimmed during replay
- During replay we trim non-auth inodes on EExport or EImportFinish abort.
Subtree trimming may be delayed, too.
Skip ... - 11:32 PM Revision c7385c1d (ceph): mds: use helpers for import_reverse
- Use helpers for common code shared between handle_export_cancel and
handle_mds_failure_or_stop.
Also include handlin... - 11:32 PM Revision e15d9ca1 (ceph): mds: cancel exports in PREPPING state on any failure
- The prepping nodes may need to discover bounds from the failed node and
may hang indefinitely. Meanwhile, we won't s... - 11:32 PM Revision 07098fa5 (ceph): mds: clean up trim_non_auth_subtree output
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:32 PM Revision d94c69e5 (ceph): mds: fix export cancel during IMPORT_PREPPING
- If we are in PREPPING, we need to drop the stickydirs() on the inodes, and
not the pins on the dirfrags. Do this in ... - 11:32 PM Revision f8551514 (ceph): mds: keep import/export subtree_map state in sync with journal
- We were being sloppy before with the ESubtreeMap vs import/export events.
Fix that by doing a few things:
- add an ... - 11:32 PM Revision 24f35e79 (ceph): mds: fix null deref in debug
- The *dir isn't always non-null (namely, during DISCOVERING state).
Signed-off-by: Sage Weil <sage@newdream.net> - 10:54 PM Revision 2c0b8440 (ceph): man: Update cmds documentation.
- You always need to specify a rank if you do journal-check.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> - 10:29 PM Revision 92d349c6 (ceph): config: warn about old-style conf section names
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:11 PM Revision ff5bc8fd (ceph): mkcephfs: fix check for highest osd
- This breaks on osd0. I was doing something stupid with sed but I can't
figure out what right now, but osdN support i... - 09:46 PM Revision e34e332b (ceph): vstart.sh: use new-style section names in config
- Use new-style section names in vstart.sh.
Also update sample.ceph.conf.
Signed-off-by: Colin McCabe <colin.mccabe@dr... - 09:40 PM Revision cc0e5668 (ceph): mon:don't check for old-style monitor section name
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:34 PM Revision a6f0ce42 (ceph): cconf: update man page
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:04 PM Revision 925a2e0e (ceph): mkcephfs, init-ceph: tolerate complete lack of a type
- We were bailing out of mkcephfs with a config with no mds's defined
(because we set -e and grep returns an error here... - 09:04 PM Revision a7e7bdee (ceph): objecter: log when we defer a write because of FULL osdmap flag
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:03 PM Revision c9154c09 (ceph): mkcephfs, init-ceph: tolerate complete lack of a type
- We were bailing out of mkcephfs with a config with no mds's defined
(because we set -e and grep returns an error here... - 09:03 PM Revision a0a5ea4e (ceph): config: do not accept old-style section names
- Stop accepting old-style section names of the form $type$id. Instead,
we want section names of the form $type.$id. ... - 08:58 PM Revision b7406701 (ceph): cconf: fix usage; clean up some code
- cconf: fix obsolete usage message. Add --list-all-sections flag.
Use new ceph_argparse stuff. Update tests.
Signed-o... - 08:18 PM Revision ba69603e (ceph): config: normalize key names, cleanup
- Normalize key names in md_config_t::get_val and md_config_t::set_val
Remove unused fields from struct config_option.... - 06:15 PM Revision 9b7f223a (ceph): rgw: fix other err related issues
- also remove the now redundant formatter->flush()
- 05:52 PM Revision 303420bf (ceph): rgw: adjustments to error handling
- fixing mixup between s3 error code and s3 error message
- 05:38 PM Revision 54bc4b69 (ceph): libceph: implement ceph_conf_set and ceph_conf_get
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:24 PM Tasks #1005 (In Progress): xlock is not unpinning during rename across MDSes
- Okay, reproduced with my extra logging. Locks 0x1163068 and 0x1162fb8 take an auth_pin but don't put one; these are t...
- 01:45 PM Tasks #1005 (Resolved): xlock is not unpinning during rename across MDSes
- See logs in kai:~gregf/logs/fsstress/freeze_tree_assert.
I managed to narrow it down to inode 20000000166 having a... - 03:22 PM CephFS Bug #1006 (Resolved): clean up cmds journal-check mode
- I'm not sure it should run in background instead of foreground.
Even if it should run in background, it should not t... - 12:18 PM CephFS Tasks #1002: Assert failure in Locker::handle_file_lock
- Yeah, Sage said this looks like the new auth is sending messages before it should, probably due to some kind of twidd...
- 11:25 AM Linux kernel client Bug #1004 (Resolved): hanging writes, bisected to i_state locking change
- ...
- 10:33 AM Feature #899 (Resolved): autotest: hammer rbd mount(s) with standard fs benchmarks
- Pushed to autotest master now.
- 04:18 AM Revision 13c6b902 (ceph): mds: init metablob MDLog* for EImportStart
- This will initialize metablob.my_offset, which makes the parent inode
journaling logic work properly.
Signed-off-by:... - 04:17 AM Revision 68653b38 (ceph): init-ceph: no log_dir default
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:17 AM Revision 6150f581 (ceph): mds: fix journal offset types
- Always uint64_t!
Signed-off-by: Sage Weil <sage@newdream.net> - 04:17 AM Revision f5b566b0 (ceph): mds: show migration state names on cancel
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:13 AM Revision b5f6eb12 (ceph): rgw: rework error handling a bit
- Rados Gateway: get rid of RGWOp::err. We already have req_state::err and
that represents the same thing.
Standardize...
04/14/2011
- 10:52 PM Revision 4eb9d4f7 (ceph): config: add test for override ordering, comment
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:49 PM Revision a1994162 (ceph): config: de-globalize reading config file
- Reading a config file into any md_config_t structure except g_conf used
to be impossible. This is because the config_... - 10:18 PM Revision 22dbbe85 (ceph): radosgw_admin: fix make check
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:14 PM Revision a3a1a043 (ceph): config: make md_config_t.name a value, not ptr
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:59 PM Revision a3144a11 (ceph): rgw: don't modify object owner when setting acls
- 09:43 PM Revision 54913d43 (ceph): radosgw_admin: add 'bucket unlink' option
- 09:43 PM Revision 6f730af1 (ceph): rgw: allow changing acl using canned acl
- 06:38 PM Revision 08e9855d (ceph): mkcephfs: Actually do a mkfs.btrfs
- Signed-off-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 05:14 PM rgw Bug #977 (Resolved): radosgw should provide error code for bad bucket name errors
- fixed by commit:b5f6eb12e1004888bf9dec6aa9c4b0499093ebfb
- 11:34 AM rgw Bug #977: radosgw should provide error code for bad bucket name errors
- An RGW error return for an illegal mkbucket:...
- 02:47 PM rgw Bug #981 (Resolved): radosgw bucket.set_acl() seems to always fail with AccessDenied
- Fixed at commit:6f730af1926f9d4fe5904ac10dbe019c0a6aef8d.
- 11:22 AM Bug #924: Various OSD/FileStore bugs
- This actually may have been fixed in e26626a8b4d91bb827a4909fd7f45f2c5f06f331.
- 11:19 AM CephFS Tasks #1002 (In Progress): Assert failure in Locker::handle_file_lock
- Looks like there's a problem because the inode in question is being renamed across MDSes (big shocker, I know!). The ...
- 11:16 AM Bug #959 (Resolved): use sizeof rather than if (0) in logging macros (and other macros)
- I checked back on this, and gcc (now) seems to be pretty smart about omitting code that's inside an if (0) { } block....
- 12:40 AM Revision 141178f1 (ceph): MonitorStore: use sync_filesystem when available
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
04/13/2011
- 10:36 PM Revision 09d4c96d (ceph): dout: log_per_instance should work with log_file
- Now log_per_instance (the symlink dance) works with both log_file and
log_dir. This will facilitate gradually removin... - 10:05 PM Revision 98911b6c (ceph): RadosModel: error handling fixes
- ReadOp should read the recieve length to prevent buffer error.
Check error codes on WriteOp and ReadOp.
Signed-off-... - 09:51 PM Revision 839b4cac (ceph): filestore: fix do_getxattr check
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:23 PM Revision 9ac7825d (ceph): FileStore: give better error message about xattrs
- Fixes #952.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 08:58 PM Revision 607ce44a (ceph): mds: do not start_new_segment on replay_start
- We do not need to start a new segment after replay. And in fact must not
journal an ESubtreeMap prior to doing resol... - 08:58 PM Revision 7dc82d09 (ceph): mds: during export list target second
- We need to maintain the invariant that (dir_auth.first==whoami) == is_auth.
Signed-off-by: Sage Weil <sage@newdream.... - 08:58 PM Revision 9e0565cd (ceph): mds: fix dn unlocking on export_reverse
- Triggered by mds_kill_import_at 5. We were clearing the export_locks
prior to calling export_unlock (der!).
Signed-... - 08:58 PM Revision 4001b81d (ceph): mds: fix choose_lock_state() on xlocked object
- This crops up on inodes during clientreplay when we reconnect the cap
on the newly created (and still xlocked) object... - 08:58 PM Revision aa074898 (ceph): mds: fix _freeze_dir assert for refragment case
- The is_freezeable_dir() is true at freeze time but not forever after over
the lifetime of the freeze. We split later... - 08:58 PM Revision 4fe48ed8 (ceph): mds: don't check_rstats on non-auth or frozen dirs
- If we are, say, auth but frozen (mid-import) the dir content isn't valid
and check_rstats will likely fail.
Signed-o... - 08:58 PM Revision c9ce4cf9 (ceph): mds: fix resolve
- This was broken by a01fba175b646f6 when an ambiguous import was changed
from CDIR_AUTH_UNKNOWN to <whoami,whoami> and... - 08:58 PM Revision 338c9a66 (ceph): mds: queue rejoin_waiters on rejoin_ack survivor
- For recovering nodes, we eventually open_snap_parents and much later
requeue these waiters. A surviving node wasn't ... - 08:58 PM Revision 7c17cd32 (ceph): mds: fix bad import_state check on handle_export_discover
- This populates import_state[] with an bad value and leads to crashes like
mds/Migrator.h: In function 'static const ... - 08:45 PM Revision e341fe0b (ceph): osd: Use syncfs when available
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:29 PM Revision 5c06fc1c (ceph): rgw: create bucket with empty name should return a valid error
- 06:52 PM Bug #1003 (Resolved): mon: health should show % degraded, unfound (and warn appropriately)
- (if it doesn't already)
- 05:39 PM Revision 59309b37 (ceph): rgw: recreation of bucket returns success
- unless it was owned by a different user, at which case it
returns -EEXIST. - 05:36 PM Revision b7b66ac9 (ceph): mds: update rstats on stray dir when you rename over existing inode.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 05:02 PM Revision cc162771 (ceph): sample.ceph.conf: add log file and pid file
- These really should be included in a sample...
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 05:01 PM Revision 694c4379 (ceph): vstart.sh: use "log file" instead of "log dir"
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:05 PM Bug #910: Multi-MDS Ceph does not pass fsstress
- Okay, been pecking away at this and I think I'm now down to one bug that Sage says he's fixed in his branch, and #100...
- 04:05 PM CephFS Tasks #1002 (Resolved): Assert failure in Locker::handle_file_lock
- ...
- 03:45 PM Revision fd0290a0 (ceph): rgw: listing non existent bucket returns NoSuchBucket
- 02:24 PM Bug #952: Make lack of btrfs/user_xattr more obvious in mkcephfs
- I'm actually glad to see a better error message go here. This was a very common problem for users!
- 02:23 PM Bug #952 (Resolved): Make lack of btrfs/user_xattr more obvious in mkcephfs
- fixed by commit:9ac7825dc1afe2e052287a669130a56cadd21af1
- 02:08 PM Bug #989 (Resolved): rstats and fragstat messages in mds logs
- Seems to be working. Pushed to master in commit:b7b66ac95e406076608df24424ac6d95d72a9ac6.
- 01:47 PM rgw Bug #985: radosgw listing keys of a bucket with name >=251 characters fails with NoSuchKey
- I added too-long-object-name error checking to rados with commit:694c4379ceb892a78a9e0f63bfe11070eff53a84 . Yehuda is...
- 01:46 PM Bug #953 (Resolved): osd, mon: use sys_syncfs() when it hits glibc
- implemented by commit:e341fe0b1854aa1d6e961395877e643a0e8b2306
- 01:27 PM rgw Bug #982: radosgw should fail with UnresolvableGrantByEmailAddress on add_email_grant to unknown ...
- Actually, looking at the code we currently just ignore entries that don't resolve. Failing with the correct error sho...
- 01:17 PM rgw Bug #983 (Resolved): radosgw hangs, gives 500 internal server
- Fixed, commit:5c06fc1c8b4e983acd7f01f7b908ba8f357621b2.
- 11:02 AM rgw Bug #999 (Resolved): Duplicate Bucket Created
- Fixed now, commit:59309b376c5356afd97a2e72f126159d2a4877ba.
- 10:29 AM rgw Bug #978 (Resolved): radosgw bucket creation should succeed on pre-existing bucket if the owner m...
- Fixed now, commit:59309b376c5356afd97a2e72f126159d2a4877ba.
- 09:22 AM Bug #1001 (Resolved): dead mds remains up, won't let others take over
- 08:34 AM rgw Bug #976 (Resolved): Non-existent bucket error code should be NoSuchBucket, not NoSuchKey
- Fixed at commit:fd0290a00ac76bbfe15bb12ee35e54dc58cf074f.
- 05:20 AM Bug #990: osd: PG::replay_queued_ops
- According to sjust this should have been fixed in the osd_wip1000 branch, but I just hit the bug again on several OSD...
- 02:57 AM Revision baeab784 (ceph): Merge remote branch 'origin/mon_mds'
- 12:20 AM Revision 02e93108 (ceph): osd: move MAX_CEPH_OBJECT_NAME_LEN into object.h
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:09 AM Revision 40828086 (ceph): osd: check obj name length to avoid ENAMETOOLONG
- Since the object store is ultimately based on ext3, ext4, or btrfs, and
object names ultimately get translated into f...
04/12/2011
- 11:40 PM Revision 2380e70a (ceph): librbd: don't write to stdout
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 09:31 PM Revision 1eae9d6d (ceph): Add test_mutate
- Add test_mutate, in an effort to track down an objecter bug.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 09:14 PM Revision 9c731edd (ceph): mdsmap: initialize standby_for_rank
- This is initialized in MDSMonitor anyway; do so where for completeness.
Signed-off-by: Sage Weil <sage.weil@dreamhos... - 09:13 PM Revision 5e27a079 (ceph): mon: simplify mds follow checks
- Instead of assigning followers in the last_beacon laggy check loop, do it
at the end, the same way we let standby nod... - 09:08 PM Revision 2b59bc61 (ceph): mon: simplify mds laggy check
- We should never have a laggy standby, so technically this doesn't change
any behavior, but it makes the flow less con... - 09:08 PM Revision e9afe17e (ceph): mon: don't take over for a standby-replay
- If a standby-replay is laggy we shouldn't "take over" for them (they're
not part of the cluster yet). They should be... - 09:06 PM Bug #989: rstats and fragstat messages in mds logs
- All right, try this patch and see if it fixes it?...
- 09:04 PM Revision 2b2766dd (ceph): rados-tool: use init_with_config interface
- Programs that use both librados and common_init should use
init_with_config.
Signed-off-by: Colin McCabe <colin.mcca... - 07:39 PM Revision 466306de (ceph): mds: make _create_system_file dirty dentries properly
- Properly dirty the new dentries so they get written to the directory
objects later on.
Signed-off-by: Sage Weil <sag... - 07:39 PM Revision d7544dd6 (ceph): mds: fix create_mydir_hierarchy to save dir
- Mark the dentries dirty so they get saved to disk (they're not journaled!).
This fixes rstat problems on startup, whe... - 06:12 PM Revision 43d74721 (ceph): mds: reset dirty->flushing on rejoin scatterflush
- Reset dirty/flushing state during rejoin.
Signed-off-by: Sage Weil <sage@newdream.net> - 06:12 PM Revision c9872a44 (ceph): mds: send any dirty scatterlock state on rejoin
- Not just inodes for auth dirfrags, but for any inode with dirty scatterlock
state. Include the root inode.
Signed-o... - 06:12 PM Revision c7f9d2f9 (ceph): mds: clear flush state on rejoin ack
- If we sent scatterlock state during rejoin, the auth will send us an inode
base. Clear scatterlock flush state if th... - 06:12 PM Revision bc98fe7e (ceph): mds: improve scatterlog debug msg
- 06:12 PM Revision 2a1af8b8 (ceph): mds: drop debugging cruft
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:12 PM Revision 0f10e954 (ceph): mds: make _create_system_file dirty dentries properly
- Properly dirty the new dentries so they get written to the directory
objects later on.
Signed-off-by: Sage Weil <sag... - 06:12 PM Revision 31fc9edb (ceph): mds: fix create_mydir_hierarchy to save dir
- Mark the dentries dirty so they get saved to disk (they're not journaled!).
This fixes rstat problems on startup, whe... - 06:06 PM Bug #997 (Won't Fix): qemu-iotest 016 fails for rbd
- We're skipping this test instead of being growable by writing to a larger offset than our size.
- 06:05 PM Bug #998 (Can't reproduce): qemu/librbd race conditon
- Stefan had no core file, and neither he nor I could subsequently reproduce this.
- 05:55 PM Revision f4d86777 (ceph): radosacl: read Ceph configuration file
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:24 PM rgw Bug #985: radosgw listing keys of a bucket with name >=251 characters fails with NoSuchKey
- this is related to #963
- 03:07 PM rgw Bug #985: radosgw listing keys of a bucket with name >=251 characters fails with NoSuchKey
- This is a problem with long pool names in the Object Storage layer....
- 04:40 PM Revision ceba04f4 (ceph): cephtool: don't print anything on SIGTERM, SIGINT
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 03:00 PM Bug #1001: dead mds remains up, won't let others take over
- mds went from up:replay to up:standby-replay:...
- 01:17 PM Bug #1001 (In Progress): dead mds remains up, won't let others take over
- 01:15 PM Bug #1001: dead mds remains up, won't let others take over
- All 3 nodes were running 0.26 plus stable patches c494689062c9, plus a patch that relaxes the journaler _trim_finish ...
- 01:09 PM Bug #1001 (Resolved): dead mds remains up, won't let others take over
- 3-node cluster, with 3 mons, 3 mdses (all configured for standby-replay), 3 osdes (but node 0 down, because I'm using...
- 01:16 PM CephFS Bug #966 (Can't reproduce): mds: scatterstat error
- Can't reproduce this one. Maybe it was resolved by commit:466306de3aeca22311993bf5a1955281499d751d.. or something ear...
- 01:06 PM CephFS Bug #933 (Can't reproduce): mds: standby-replay incorrectly probed journal for live mds
- i give up! maybe this was triggered prior to the WR flag being set during filer probe.
- 11:38 AM Bug #1000 (Closed): osd: PG::replay_queued_ops
- Uh, my bad, this is a duplicate of #990.
- 10:16 AM Bug #1000: osd: PG::replay_queued_ops
- Likely caused by handle_pg_notify calling do_peer on an active pg.
- 10:01 AM Bug #1000 (Closed): osd: PG::replay_queued_ops
- On my cluster (40 OSD's) I'm seeing multiple OSD's going down with this backtrace:...
- 11:34 AM Bug #802 (Closed): osd: failing to send heartbeats (btrfs hang?)
- We're chalking this up to insufficient CPU to handle all of the threads. There are some tricks we could play (bumpin...
- 11:33 AM CephFS Bug #637 (Resolved): mds: check replica scatterlock flush on rejoin
- fixed by commit:c7f9d2f9a8607d1147c4ee49f8fb565fde2f3278 and prior commits.
- 11:23 AM rbd Bug #944 (Duplicate): rbd snapshots not exposed as read-only
- Forgot about this ticket - it was fixed in #927.
- 10:35 AM rgw Bug #983: radosgw hangs, gives 500 internal server
- I reproduced this. I'm still trying to figure out why this happened. Best guess so far is a timeout on the Rados end ...
- 09:47 AM Bug #993 (Resolved): ceph -w whines on quit
- Resolved by commit:ceba04f4dde5b6538ddc648af740575090a9bf60
- 09:45 AM Bug #881 (Resolved): confutils: memory leaks
- I don't see any obvious way to get valgrind to stop warning about these "potential leaks" which are not leaks at all....
- 04:09 AM Revision 21df0427 (ceph): mds: don't crash on bad export_dir command
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:09 AM Revision fb8c09b0 (ceph): mds: fix file_eval on subtree roots
- We were going from mix->sync when we shouldn't. Fix mis-coded filelock
and look at _lock_ state, not inode's!
Signe...
04/11/2011
- 11:30 PM Revision 1f23a229 (ceph): Merge remote branch 'origin/dead-code'
- 11:24 PM Revision 60b645e5 (ceph): osd: fix race in op enqueuing
- This is never really exposed except when op_threads == 0.
Signed-off-by: Sage Weil <sage@newdream.net> - 11:24 PM Revision 6768586c (ceph): osd: take reference for RepModify
- We put in the completions on done; take the matching reference.
Signed-off-by: Sage Weil <sage@newdream.net> - 11:24 PM Revision 7ca69ebe (ceph): osd: fix ref leak on pg when queueing
- op_wq.queue() takes a ref; don't take one here too. We were leaking like
crazy.
Signed-off-by: Sage Weil <sage@newd... - 09:23 PM rbd Bug #944: rbd snapshots not exposed as read-only
- was this fixed already? i forget!
- 09:23 PM Tasks #895 (Resolved): make backlogs charts work
- 09:18 PM rgw Bug #999: Duplicate Bucket Created
- i recently changed the creation code some. and see also #978.
- 05:38 PM rgw Bug #999 (Resolved): Duplicate Bucket Created
- I think I managed to created a bucket called 'foobuckat' on both my account (webid matwod) and Sage's account. I thi...
- 06:43 PM Bug #989: rstats and fragstat messages in mds logs
- The log file up until a ways after the first check_rstats problem was just over 5MB, so I split it up here such that ...
- 11:35 AM Bug #989: rstats and fragstat messages in mds logs
- Okay. An old filesystem wouldn't have explained the negative rstats anyway.
Let me know what you come up with! - 11:12 AM Bug #989: rstats and fragstat messages in mds logs
- Every test I run is on a freshly installed cluster.
We've automated all the underlying stuff such that it's pretty q... - 11:09 AM Bug #989: rstats and fragstat messages in mds logs
- Well the MDS directory is a permanent thing -- once it's broken it's going to stay broken. Ideally we'd reproduce wit...
- 10:59 AM Bug #989: rstats and fragstat messages in mds logs
- This may be something popping out of the MDS changes intended for NFS support (inode lookup).
I'm going to retest by... - 09:11 AM Bug #989: rstats and fragstat messages in mds logs
- The more I look at this the less sense it makes. The only things in the mds dir are the journal and the stray dirs. Y...
- 05:37 PM Bug #993: ceph -w whines on quit
- I guess it's ugly that it echoes it twice. In general, we probably don't need to blast the message out to every possi...
- 04:45 PM Revision eacfb8d3 (ceph): TestSnaps: Fix compile error
- Introduced in 0e371de79e4047b92317dee353f6d75c2aba4260.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 04:43 PM Revision cc5dd6d6 (ceph): mon: simplify osd health output
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:36 PM Revision 9c0d42fe (ceph): osd: only log scrub errors (not ok's)
- Otherwise we fill up the error log with "foo is ok!" messages.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:22 PM Linux kernel client Feature #448 (Rejected): support dns resolution in libceph
- 03:22 PM Linux kernel client Feature #448: support dns resolution in libceph
- Indeed, that's the plan: #958
- 03:15 PM Linux kernel client Feature #448: support dns resolution in libceph
- How about writing a userspace wrapper for that echo >/sys, instead? There's no reason for this functionality to be in...
- 03:13 PM Bug #998 (Resolved): qemu/librbd race conditon
- As reported in http://marc.info/?l=kvm&m=130225223223016&w=2:
There's a race condition that sometimes causes aio_r... - 03:11 PM Bug #997 (Won't Fix): qemu-iotest 016 fails for rbd
- This test treats an rbd image as a 'growable' block device, which will accept reads and writes at large offsets and r...
- 02:43 PM Feature #745 (Resolved): qa: radosgw tests
- 01:19 PM Bug #906: clustered mds: lchown not setting uid/gid
- Re-running as job 409, clocks are in decent sync:...
- 11:42 AM Bug #931 (Resolved): opening up security. xattr
- Believe the patch from last week fixed this.
04/10/2011
- 09:28 AM Bug #996: osd: ReplicatedPG::do_osd_ops
- The OSD code was briefly broken last week -- it looks like you managed to pull down a bad one. Get the latest master ...
- 01:32 AM Bug #996 (Can't reproduce): osd: ReplicatedPG::do_osd_ops
- I had an issue that my cluster would not recover, spoke with joshd about this on IRC ( http://irclogs.ceph.widodh.nl/...
04/09/2011
- 01:16 AM Revision 1cb4aae4 (ceph): ceph_fuse_ll_main: restructure error handling
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 01:16 AM Revision c02b56e6 (ceph): cfuse: pass through SIGTERM, SIGINT to FUSE
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:39 AM Revision 46c761fa (ceph): librbd: check for out of bounds I/O in all forms of read/write
- This was found by qemu-io tests, which tried to read and write past
the end of an rbd image. The test hung waiting fo... - 12:39 AM Revision b3f5102d (ceph): librbdtest: check for error on out of bounds I/O
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
04/08/2011
- 11:15 PM Revision 0e371de7 (ceph): TestSnaps.cc: allow client id to be overridden
- CEPH_CLIENT_ID environment variable if set now overrides the rados
client id for testsnaps
Signed-off-by: Samuel Jus... - 11:12 PM Revision e34078b2 (ceph): cfuse recognizing same xattr subset as kernel client
- Hopefully this patch is formatted/signed off correctly.
Signed-off-by: Brian Chrisman <brchrisman@gmail.com> - 08:11 PM Revision 2b0e474a (ceph): rgw: utility, not daemon
- - we don't want to open a log file
- we don't want to print a stability warning
- we don't want to fork (not that we ... - 06:17 PM Revision 7f3a0540 (ceph): CDir: check_rstats will now print out dir stats whenever there's a bug.
- Previously it only printed out dir stats at high debug levels, which
meant you could get output of the dentries witho... - 04:26 PM Bug #989: rstats and fragstat messages in mds logs
- Full section of othat log
- 04:17 PM Bug #989: rstats and fragstat messages in mds logs
- I see more debug messages here now. This looks like one cycle of it, but I can turn my debugging up higher if necess...
- 10:53 AM Bug #989: rstats and fragstat messages in mds logs
- I actually just pushed a slightly cleaner one to master so if you're pulling on a regular basis no need to patch. :)
... - 10:47 AM Bug #989: rstats and fragstat messages in mds logs
- I'll test this... the bigger issue I've found is that the actual problem I was investigating when I saw these log fil...
- 09:12 AM Bug #989: rstats and fragstat messages in mds logs
- ...
- 02:30 PM Bug #903 (Resolved): write an autotest that calls test-pybind.py
- implemented in commit:cb85b4db7bd708351ebab5b385419b09413469dd
- 11:30 AM Feature #951 (Resolved): set up semi-permanent cosd cluster, dev backups
- 11:23 AM CephFS Bug #994: EOpen reply on non-auth MDS is busted
- Ooof, not a simple thing at all:
1) The crashing MDS is non-auth.
2) That means the directory gets trimmed during r... - 10:20 AM CephFS Bug #994 (In Progress): EOpen reply on non-auth MDS is busted
- Hopefully I can figure this out for .27 -- replay problems are more important than the horde of multi-MDS stuff uncov...
- 10:19 AM CephFS Tasks #973 (Resolved): Dir failing to freeze
- Appears to be working.
- 10:18 AM CephFS Tasks #922 (Resolved): fsstress: Request ping-pongs when dentry and inode auth are separate
- Appears to be working
- 09:51 AM Linux kernel client Bug #852: secret key shows up in /etc/mtab after mount -o secretfile=/.../key
- It needs commit 4b2a58abd1e17c0ee53c8dded879e015917cca67 on the kernel side, first included in v2.6.39-rc2.
- 09:15 AM Linux kernel client Bug #852: secret key shows up in /etc/mtab after mount -o secretfile=/.../key
- Thanks! It seems that this fix missed ceph-0.26, even though mount.ceph (that presumably was the bit that needed fix...
- 09:48 AM Bug #995 (Can't reproduce): unrecoverable 0.*p* PGs
- Several times, with 0.25.1, presumably because of random btrfs hangs (now strongly linked with an active local ceph m...
- 09:34 AM Bug #943 (Won't Fix): 3-mon cluster won't start
- Ah, that makes sense. I think we'll drop this one unless somebody with a more likely network configuration runs into ...
- 09:29 AM Bug #943: 3-mon cluster won't start
- 'fraid I no longer have that log file :-(
From what I remember, it had lots of messages from a bunch of osds each ... - 09:22 AM Bug #893: no filesystem created if all mdses are configured for standby-replay
- Can you post the backtrace? We fixed a few bugs with standby-replay in the master branch already.
- 09:09 AM Bug #893: no filesystem created if all mdses are configured for standby-replay
- 'fraid this didn't quite work. Just tried creating a new filesystem with 0.26. An MDS marked as standby-reply does ...
- 03:07 AM Revision f3f7383c (ceph): obsync: preserve ACLs
- Preserve ACLs by default. Also add --no-preserve-acls option to disable
this behavior.
Signed-off-by: Colin McCabe <...
04/07/2011
- 11:30 PM Revision 49951173 (ceph): osd: futz with stat again.
- We really need to fill in the result, even if it's that the object
doesn't exist!
Signed-off-by: Greg Farnum <gregor... - 11:09 PM Revision 556f06e8 (ceph): obsync: escape obj names between S3 and local FS
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:08 PM Revision 0de233c3 (ceph): test/osd: Add latency stat collection to TestSnaps.cc/RadosModel.h
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 11:07 PM Revision e26626a8 (ceph): ReplicatedPG: fix error in snap_trimmer
- newsnaps continued to be used after it was swapped with coi.snaps.
Signed-off-by: Samuel Just <samuel.just@dreamhost... - 10:05 PM Revision 84aaa6e0 (ceph): crush, clitest: Port old test_crushtool.sh to clitest.
- Also test --help, while we're at it.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 10:01 PM Revision 61ce52cc (ceph): osd: more futzing with stat
- We can get here when the object doesn't exist if the client specifies
may read and may write (in this case, Filer::pr... - 09:53 PM Revision eb08daab (ceph): libceph: Make header usable from plain old C.
- Fixes: #986
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 09:42 PM Revision eee1f920 (ceph): crush: Remove dead code.
- The Makefile never integrated with the surrounding automake
setup, and is unused.
The test helper has failed to comp... - 09:37 PM Revision 9599b66f (ceph): ReplicatedPG: tolerate ENOENT when statting.
- This got broken in commit f842bcf893e42a4cefda07d0bbf86303b17ddcd7
so that on any MDS restart the OSD would die. Bad ... - 09:35 PM Revision c4946890 (ceph): Don't crash during initialization when working without an osd journal.
- Signed-off-by: Alexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: Sage Weil <sage@newdream.net> - 09:34 PM Revision aefa1046 (ceph): crush: Remove dead code.
- Looks like these never went anywhere.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 09:32 PM Revision 38c6831c (ceph): radosgw: Remove PHP setting from VirtualHost documentation
- Signed-off-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Sage Weil <sage@newdream.net> - 09:25 PM Revision 05596d7e (ceph): osd: Remove dead code.
- RAID4PG has been "temporarily ripped out" for 3.5 years.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 09:22 PM Revision f81b505c (ceph): msg, os: Remove dead code.
- Looks like these never went anywhere.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 09:22 PM Revision 42975b12 (ceph): obsync: add --follow-symlinks
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:22 PM Revision 26533610 (ceph): boto_tool.py: fix rmbucket
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:22 PM Revision 2506819d (ceph): test-obsync: some fixes, be more verbose
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:22 PM Revision 033e29a4 (ceph): test-obsync.py: add test for --follow-symlinks
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:07 PM Revision a66da4af (ceph): qa: Remove fsx workunit.
- Covered better by autotest, these days.
- 09:06 PM Revision d55e98b5 (ceph): common, msg: Remove dead code.
- Looks like these never went anywhere.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 08:45 PM Revision 4537d92d (ceph): crush.old: Remove dead code.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 08:07 PM Bug #956 (Resolved): obsync: preserve ACLs when doing bucket to bucket transfers
- implemented by commit:f3f7383ce98821dfb0c1c8cc1140a9576e51c25c
- 07:40 PM Revision e01cd3b5 (ceph): ebofs: Remove dead code.
- Nothing has used the EboFS object storage in years.
Fix usage string of dupstore.
FileJournal and Journal are now p... - 07:28 PM Bug #989: rstats and fragstat messages in mds logs
- That excerpt was basically where the log first started going into the rstats dump output.
I'll grab the commit id...... - 04:42 PM Bug #989: rstats and fragstat messages in mds logs
- Well, I was unable to reproduce #966. Maybe Sage can at some point, but in the meantime let's look at this one. You'r...
- 10:19 AM Bug #989: rstats and fragstat messages in mds logs
- Snagging this on the assumption it's the same as #966, which I'll be looking at today.
- 07:24 PM Revision b293d451 (ceph): osbdb: Remove dead code.
- Nothing has used the BerkeleyDB-backed object storage in years.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamh... - 07:03 PM Revision 78716a4c (ceph): clitest: Disabled problematic rados test, for now.
- Failed on gitbuilder due to odd libtool interaction.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 06:12 PM Revision fe157279 (ceph): clitest: Add more simple basic functionality tests.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:58 PM Revision 8c30f535 (ceph): librados-config: Distribute the librados-config.8 manpage.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision 456f2fc5 (ceph): pybind: Distribute the test script, too. Fixes "make distcheck".
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision 3a80cc62 (ceph): buffer: Move buffer accounting into functions, don't expose atomic_t.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision 48785750 (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision 14cff958 (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision f0f4b80f (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision a752d83e (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision 2bd9ce11 (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision ce767984 (ceph): buffer: Move functions touching buffer::raw internals into buffer.cc
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision c26158bf (ceph): buffer: Move buffer::raw* out of the header.
- This paves the way for getting librados.hpp clean of atomic_t
leaking to client code.
Signed-off-by: Tommi Virtanen ... - 05:53 PM Revision 699c6949 (ceph): buffer, atomic: Explicitly include assert.h where asserts are used.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:53 PM Revision bcacb137 (ceph): buffer: Move atomic.h include from buffer.h to buffer.cc
- Fix up other places that failed to include atomic.h even
if they used atomic_t directly.
Fixes: #979
Signed-off-by: ... - 05:53 PM Revision 536ae0ca (ceph): librados: Stop bundling the now-unused headers.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:04 PM Bug #935: ceph client heavy loading in Metadata server
- Your pictures have cut off the warning messages -- can we get images with the whole screen, please? :)
You're usin... - 04:57 PM Bug #943: 3-mon cluster won't start
- Sorry Alexandre, I guess we lost this in the shuffle.
A programmatic solution based on log size is certainly possi... - 04:26 PM CephFS Bug #994 (Resolved): EOpen reply on non-auth MDS is busted
- Saw this while trying to reproduce #966. After restarting the MDSes, one of them crashed with:...
- 04:09 PM Bug #941 (Resolved): obsync: escape object names that end in slash
- implemented by 556f06e8d1de1c46fe02c511b35b54a0009eb2f7
- 04:04 PM CephFS Bug #966: mds: scatterstat error
- I haven't been able to reproduce this even once. Assigning back to Sage per his request.
- 01:41 PM CephFS Bug #966: mds: scatterstat error
- Oh, that assert actually occurred because of an ENOENT return code, not a stat mismatch. Pushed a fix, will try again.
- 11:35 AM CephFS Bug #966: mds: scatterstat error
- Ugh. My first attempt to reproduce this, the OSD crashed before the MDS did:...
- 10:19 AM CephFS Bug #966 (In Progress): mds: scatterstat error
- I'll look into this today.
- 10:12 AM CephFS Bug #966: mds: scatterstat error
- Update this item when there's a fix submitted and I can retest my related issue fairly quickly.
- 02:54 PM CephFS Bug #986 (Resolved): libceph is not safe to call from C
- Pushed to master as eb08daaba28805c24bba654559a06983779ed8a6.
- 02:48 PM CephFS Bug #986: libceph is not safe to call from C
- sounds good!
- 01:19 PM CephFS Bug #986: libceph is not safe to call from C
- Looks good. It was supposed to be C-happy, but I was even less familiar with C back then than I am now so there were ...
- 11:49 AM CephFS Bug #986 (In Progress): libceph is not safe to call from C
- Actually, all it needs is commit dd9ffc080f73fdb18f7919878f2ddf7e47f7fa3f, in branch libceph-from-c. The header uses ...
- 01:52 PM Bug #993 (Resolved): ceph -w whines on quit
- ...
- 11:01 AM Feature #745 (In Progress): qa: radosgw tests
- The s3-tests.git repository is looking good, a bunch of tickets filed, two things missing: autotest automation (to be...
- 11:00 AM Bug #979 (Resolved): 3rd party compilation, e.g. against librados, is broken
- 10:20 AM Bug #910: Multi-MDS Ceph does not pass fsstress
- I pushed a bunch of my work on this in, but am dropping work on this for a bit to look at the rstat issue.
- 09:43 AM Bug #987 (Duplicate): librados.hpp api is not usable out-of-tree
- Whoops, this is same as #979
- 09:34 AM Bug #972 (Duplicate): Build problem with librados.hpp
- This is a duplicate, or at least part of, #979.
- 09:16 AM CephFS Tasks #973: Dir failing to freeze
- Think we got this but not sure...testing will help tell!
- 03:50 AM Bug #992: osd: OSD::_lookup_lock_pg
- The title had to be: "osd: OSD::_lookup_lock_pg", pressed the submit button a bit to fast
- 03:34 AM Bug #992 (Closed): osd: OSD::_lookup_lock_pg
- When testing with recovery of a failed machine, I encountered a few OSD crashes.
Both osd5 and osd11 went down wit... - 01:14 AM Bug #991 (Closed): osd: OSD::_process_pg_info
- At the same time my OSD from #990 crashed, another OSD on the same machine went down, but with a different backtrace:...
- 01:13 AM Bug #990 (Closed): osd: PG::replay_queued_ops
- I upgraded my cluster to commit:24caedc8f549eeeba48b2d4a44927ee16e65c42a after doing so, one of my OSD's crashed with...
04/06/2011
- 11:34 PM Revision e5fce7b8 (ceph): Revert "mds: rip out rename linkmerge support"
- This reverts commit d467fbfd1714ef04faf41e604d6a78e33ef7fa10.
The linkmerge stuff _is_ needed for reintegrating mult... - 07:35 PM Revision 7470cc4e (ceph): config: complain if --name gives an invalid type
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:51 PM Revision cbb4861b (ceph): ConfFile: add support for backslashes
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:43 PM Bug #989: rstats and fragstat messages in mds logs
- This may be a dup of #966.. we may want to fix that one first (it's easy to reproduce).
- 04:23 PM Bug #989 (Resolved): rstats and fragstat messages in mds logs
- I'd like to track down what's going wrong here. These messages generate about 500MB of mds log file while I'm copyin...
- 12:51 PM Feature #961 (Resolved): obsync: -n mode
- already implemented as -n. You can also give --dry-run.
- 12:35 PM Bug #969 (Resolved): config: --name sdfsdf.sdfsdf should fail
- fixed by commit:7470cc4eb5c399f92e5f00c3e7abe536d47dce47
- 12:03 PM Bug #512 (Resolved): rados_initialize returns 0 when ceph.conf contains no monitors
- This is fixed by the Rados API change that we did a while back. rados_initialise has been split into rados_create and...
- 12:45 AM Revision 24caedc8 (ceph): pybind: install ceph-pybind-test on debug builds
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
04/05/2011
- 10:18 PM Revision 31496495 (ceph): rgw: fix decl vs goto error
- 09:41 PM Revision 64185b68 (ceph): conf: small code cleanups
- md_config_t: store a ConfFile by value rather than by pointer.
ConfFile::parse_file: negate error when returning err... - 09:39 PM Revision 2dcfe214 (ceph): osd: fix signed/unsigned comp
- warning: osd/ReplicatedPG.cc:1208: comparison between signed and unsigned integer expressions
Signed-off-by: Sage We... - 09:12 PM Revision f3a00ce6 (ceph): ConfUtils: new parser
- A new parser for ConfUtils that outputs a list of syntax errors when
files are wrong.
It also checks that configurat... - 07:36 PM Revision f842bcf8 (ceph): osd: simplify stat
- We don't actually need to stat the object (again) since we have the size
(and mtime) in the object_info_t. Keep the ... - 06:57 PM Revision b0f817ac (ceph): osd: process missing when log is empty
- There are important cases where the replica will send a missing set and
empty log to the primary during peer (e.g., w... - 05:10 PM Bug #979 (In Progress): 3rd party compilation, e.g. against librados, is broken
- I have a version that makes buffer::raw private working, will push soon.
- 10:09 AM Bug #979 (Resolved): 3rd party compilation, e.g. against librados, is broken
- 1. include files aren't prepared to handle not getting -D__CEPH__, will do unpredictable things
2. not sure what s... - 04:38 PM Feature #988 (Duplicate): librbd: trivial layering
- Layer a writeable image over a read-only image.
See http://marc.info/?l=ceph-devel&m=129867273303846&w=2 - 04:36 PM Feature #183: qa: xfstests workunit
- xfstests includes things like fsstress. this will cover many bases at once.
- 03:36 PM CephFS Tasks #973: Dir failing to freeze
- As I said, the inode in question was migrated, so the xlock cleanup isn't going through the normal unlock paths, that...
- 12:11 PM CephFS Tasks #973: Dir failing to freeze
- This is because XLOCKDONE is unstable, and LOCK, EXCL, etc. are stable. So, on xlock, we only add a new auth_pin if ...
- 03:18 PM Bug #987 (Duplicate): librados.hpp api is not usable out-of-tree
- It exposes bufferlists which expose atomic_t internals, which are not guaranteed stable.
Discussion ongoing.
The ... - 02:40 PM Bug #881: confutils: memory leaks
- I re-ran this on the latest code, and I saw a bunch of "potential leaks" related to allocating std::strings in the Co...
- 02:17 PM Bug #904 (Resolved): confutils: should warn about syntax errors
- implemented by commit:f3a00ce6107abffd1c5396b028444dc9a33963e1
- 02:08 PM CephFS Bug #986: libceph is not safe to call from C
- See #813. We need to break this into a .h and .hpp like everything else.
- 01:36 PM CephFS Bug #986 (Resolved): libceph is not safe to call from C
- src/client/libceph.h looks like it tries to be C-compatible (extern "C" used), but it's not.
Wrapping the extern "C"... - 12:37 PM CephFS Bug #933: mds: standby-replay incorrectly probed journal for live mds
- Can't reproduce. Cleaned up stat, print full debug info.
- 11:58 AM Bug #967 (Resolved): osd: PG::do_peer crash when restarting other OSD in PG
- When I went to look at this the pg was waiting for log+missing on osd0. There was no logging, so I restarted osd0 wi...
- 11:03 AM rgw Bug #985 (Resolved): radosgw listing keys of a bucket with name >=251 characters fails with NoSuc...
- from s3-tests:...
- 10:48 AM rgw Feature #984 (New): rgw: user logging API
- from s3-tests:...
- 10:29 AM rgw Bug #983 (Resolved): radosgw hangs, gives 500 internal server
- ...
- 10:22 AM rgw Bug #982 (Resolved): radosgw should fail with UnresolvableGrantByEmailAddress on add_email_grant ...
- from s3-tests, radosgw gives 403 error, aws has a better & more explicit error:...
- 10:19 AM rgw Bug #981 (Resolved): radosgw bucket.set_acl() seems to always fail with AccessDenied
- from s3-tests, several tests fail because of this:...
- 10:16 AM rgw Bug #980 (Resolved): radosgw returns access_key instead of user_id when fetching acl
- from s3-tests: bucket.get_acl() on AWS returns the user_id (typically a 64-char hexstring), not the access_key
<pr... - 04:10 AM Revision 0139be7e (ceph): librbd: remove unnecessary includes
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
04/04/2011
- 09:57 PM CephFS Tasks #973: Dir failing to freeze
- Hmm, the locking surrounding this is definitely suspect. I'll try and hack something together, but simple_xlock code ...
- 07:03 PM CephFS Tasks #973: Dir failing to freeze
- I am reasonably certain the problem is this, from Locker::simple_xlock:...
- 06:07 PM CephFS Tasks #973: Dir failing to freeze
- Oh, suppose I should update. It's removing the direct inode pin from 10000000075 but not the one from the dir. That's...
- 09:37 AM CephFS Tasks #973 (Resolved): Dir failing to freeze
- Dir 10000000074 has got an auth_pin that won't go away. I think the problem is here:...
- 09:00 PM Revision b6084cf9 (ceph): config: expand metavariables when needed
- md_config_t::get_val_from_config_file now has an option that causes it
to expand metavariables. This is necessary bec... - 09:00 PM Revision 94fade29 (ceph): config: add ability to complain about parse errs
- Change the ConfUtils interface so that we have a way to return
information about parse errors.
Signed-off-by: Colin ... - 09:00 PM Revision 4a27cec8 (ceph): confutils: test unicode parsing
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:51 PM Revision 0e26ece4 (ceph): config: fix metavariable substitution
- common_init: Do metavariable expansion as a separate step after the
configuration, argv, and env have been read. This... - 05:14 PM rgw Bug #978 (Resolved): radosgw bucket creation should succeed on pre-existing bucket if the owner m...
- s3 test:...
- 05:10 PM rgw Bug #977 (Resolved): radosgw should provide error code for bad bucket name errors
- s3 test...
- 05:06 PM Revision 32e422e4 (ceph): configure: change gtk dep to 2.12 instead of 2.13 for lenny
- 05:06 PM Revision 9981ff90 (ceph): v0.26
- 05:04 PM rgw Bug #976 (Resolved): Non-existent bucket error code should be NoSuchBucket, not NoSuchKey
- s3 test:...
- 04:25 PM Revision 32274050 (ceph): config: remove some unecessary g_conf references
- Class methods of md_config_t should refer to the enclosing config.
Signed-off-by: Colin McCabe <colin.mccabe@dreamho... - 04:20 PM Feature #975 (Resolved): osd: use fallocate(FL_PUNCH_HOLE) for zero
- use the new FITRIM ioctl to zero an object range.
- 02:34 PM Bug #971 (Resolved): OSD's write pid to /var/run/ceph/$type.$id.pid
- confirmed fixed.
cconf is fixed too. - 01:10 PM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- did this break cconf? http://ceph.newdream.net/gitbuilder/log.cgi?log=0e26ece4e366972cbcbaf76db75df8d4512e361e
- 12:15 PM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- Yes, that works for me. PID's are correct now.
- 11:58 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- This should be fixed by commit:0e26ece4e366972cbcbaf76db75df8d4512e361e
- 11:23 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- I think I know what's going on here. Will fix in a little bit.
- 10:30 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- Just tried with commit:d94142221153ec985c699ad69c3925136f3a30de still the same result.
My ceph.conf is attached - 09:39 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- I'm afraid I'm not able to reproduce this in master. Can you try again with head-of-line and see if it is fixed there...
- 09:21 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- This doesn't occur for me at change 6966c3eda74064e766e22c21203ea30f97910f32. It is not in 0.26.
- 08:32 AM Bug #971: OSD's write pid to /var/run/ceph/$type.$id.pid
- Colin, can you verify whether this affects 0.26 too before I tag that release today? Thanks!
- 04:08 AM Bug #971 (Resolved): OSD's write pid to /var/run/ceph/$type.$id.pid
- It seems that commit:9cde1f44a57cd630a66e77ed3d59127d91834d09 has broken the writing of pidfiles for OSD's, but it co...
- 12:56 PM Feature #974 (Resolved): autotest: include ffsb benchmarks
- see http://btrfs.boxacle.net/
fwiw the one Ted T'so was running last fall was random_write.32.ffsb - 09:06 AM Bug #972 (Duplicate): Build problem with librados.hpp
- From the list:
>I'm in the progress of evaluating librados as an object store. I'm using Debian's latest packages as...
04/03/2011
- 09:07 PM Linux kernel client Bug #970: Kernel crash (cause?: lots of small files)
- Well the first line there is "can't allocate 512 bytes"...looks like you ran out of memory. Was there memory pressure...
- 07:43 PM Linux kernel client Bug #970 (Resolved): Kernel crash (cause?: lots of small files)
- ceph: 0.25.2
client: 2.6.38.1-2
default setups, 2 nodes 3 osd each
while (copying) writing a lot of files (e.g.,... - 05:18 AM Revision 05c281bf (ceph): Revert "autoconf: Complain if tcmalloc is not found."
- This reverts commit a9afdca18e2264fff70b5aaf864ae9abb0436dca.
This breaks debian build on lenny because tcmalloc isn...
04/02/2011
- 01:58 AM Bug #967: osd: PG::do_peer crash when restarting other OSD in PG
- Ok, it took a while to get there, but after about 45 mins it got to:...
04/01/2011
- 11:59 PM Revision d9414222 (ceph): cclass: add bindir option
- Ceph binaries don't live in /usr/local on autotest machines.
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> - 11:28 PM Revision 9ac44d4a (ceph): MDS: change messenger name for replay mdses
- This will cause read operations from standby mdses to be distinguishable
from those from the normal by changing the n... - 11:21 PM Revision 553da45c (ceph): mds: fix trim_non_auth assert for base inodes
- No parent dentry for base dirfrags (e.g., root).
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 11:21 PM Revision 3ce9534d (ceph): client: close snaprealm on last cap removal during reconnect
- The invariant is is_any_caps() IFF in->snaprealm, where is_any_caps() is
!caps.empty || exporting_mds >= 0. If we cl... - 10:58 PM Revision b21fbab1 (ceph): journaler: adjusted requested/received_pos when dropping journal tail
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:58 PM Revision 64afb9d7 (ceph): confutils: more parsing tests
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:57 PM Revision d4c4b4a0 (ceph): journaler: fix partial tail entry correction
- If we encounter a partial tail entry, we drop it by moving the write_pos
(end of journal) back to read_pos. We also ... - 10:34 PM Revision 2630ecc6 (ceph): config: Add confutils unit test
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:11 PM Revision 8ea93800 (ceph): journaler: fix signedness
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:09 PM Revision bb814758 (ceph): journaler: fix warning
- warning: osdc/Journaler.cc:682: comparison between signed and unsigned integer expressions
Signed-off-by: Sage Weil ... - 09:34 PM Revision 6fcecbc1 (ceph): mds: close exported dirfrag
- We have to keep export bounds open for auth subtrees. After we export a
subtree, though, there are two opportunities... - 09:34 PM Revision a47c2250 (ceph): mds: use try_trim_non_auth_subtree helper
- This helper captures the logic of keeping subtrees when necessary but
dropping them when possible, and cleaning up as... - 09:34 PM Revision a01fba17 (ceph): mds: set dir_auth between EImport{Start,Finish}
- This ensures that try_trim_non_auth_subtree() doesn't throw out a subtree
we're in the midst of importing (during jou... - 09:34 PM Revision 03c9365c (ceph): mds: rejoin dirfrags explicitly
- We need to rejoin dirfrag replicas explicitly. We were doing only dentries,
but that won't necessarily include every... - 09:34 PM Revision bafff042 (ceph): mds: fix bounds on import
- The add_ambiguous_import() call was clobbering the bounds field for
EImportStart::replay(), screwing up the subtree a... - 09:24 PM Revision 5d998a7a (ceph): MDSMonitor: MDS do not standby replay until other is active
- Previously, an mds could go into standby replay before the mds which it
is replaying has finished creating.
Signed-o... - 09:01 PM Revision ff2c2e31 (ceph): journaler: fix requested_pos badness in _issue_read
- We want to advance requested_pos *only* as far as we actually want to read,
i.e., the previously-probed end of the jo... - 07:19 PM Revision 65714c59 (ceph): mds: show correct incarnation on mdsmap update
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:48 PM Revision 1696388e (ceph): Merge remote branch 'origin/rgw'
- 06:37 PM Revision b60a3c99 (ceph): boto_tool: often use get_bucket rather than lookup
- get_bucket throws an exception when the bucket doesn't exist, whereas
lookup just returns None. Sometimes we want the... - 06:37 PM Revision dec5b6b8 (ceph): mds: fix discover_path
- If we have the base dirfrag, do not request it. Otherwise we can get a
reply that contains only that (partial progre... - 06:35 PM Revision 0fb02b25 (ceph): Merge branch 'mds_ino'
- 06:35 PM Revision 06ed898d (ceph): mds: fix find_ino_dir completion
- Pass the path back in so that we can retraverse the path on retry.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 06:24 PM Revision 2f82b394 (ceph): client: fix lookup_hash request hash value
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:24 PM Revision 8408f83c (ceph): client: lookup_ino command
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:24 PM Revision 3bfc468d (ceph): mds: don't query self from find_ino_peers
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:24 PM Revision 9bed4e3a (ceph): mds: fix discover_path
- If we have the base dirfrag, do not request it. Otherwise we can get a
reply that contains only that (partial progre... - 06:24 PM Revision 2d926968 (ceph): mds: fix dir fetch on lookup_hash fix
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:24 PM Revision c736d7f6 (ceph): mds: find_ino_dir debugging
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:00 PM Revision a603acc1 (ceph): ceph_argparse: fix silly usage message
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:27 PM Revision 760bc01f (ceph): config: typecheck initialization macros
- When initializing the config_options array, complain if the size of the
option field we're trying to initialize doesn... - 05:27 PM Revision ee483259 (ceph): config: introduce OPT_U64
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:35 PM CephFS Bug #932 (Closed): mds: standby-replay osd ops have bad entity_name_t
- 04:28 PM Revision 80d5873c (ceph): client: fix lookuphash to use rjenkins
- This is the default. Not ideal, but avoids explicitly specifying the hash
function.
Signed-off-by: Sage Weil <sage.... - 04:09 PM Revision 6b4fb92a (ceph): mds: add LOOKUPINO op
- This just looks up a single ino (instead of the ino + dirino/hash
LOOKUPHASH business).
Signed-off-by: Sage Weil <sa... - 04:06 PM Revision 1ff3ce1f (ceph): mds: rename chained _lookup_hash_? methods
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:59 PM Revision 80b621ba (ceph): mds: also try to look up ino directly
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:27 PM Bug #967 (In Progress): osd: PG::do_peer crash when restarting other OSD in PG
- First I saw 1.7a8 not peering because osd.16 wasn't sending a response. I restarted osd.16 with logging enabled... 1...
- 08:55 AM Bug #967 (Resolved): osd: PG::do_peer crash when restarting other OSD in PG
- This morning I pulled out the machine hosting osd20, 21, 22 and 23. After bringing this machine back, 3 PG's stayed i...
- 02:41 PM CephFS Bug #968 (Resolved): mds: verify_bounds error
- fixed with partial merge of mds_frag rejoin work.
- 12:09 PM CephFS Bug #968 (In Progress): mds: verify_bounds error
- 09:23 AM CephFS Bug #968 (Resolved): mds: verify_bounds error
- out.subtree on flak...
- 02:35 PM Bug #953: osd, mon: use sys_syncfs() when it hits glibc
- This is in glibc.git; will be in the next release (2.14).
int syncfs (int __fd); - 02:09 PM rgw Feature #787: rgw watchdog
- Some thoughts:
1. We probably want to support a get_cluster_status operation that gives information about the cluste... - 01:05 PM Linux kernel client Bug #915 (Resolved): kclient corrupts on direct_io? ceph_direct_io_test reliably hoses the client...
- Whoops, forgot to actually mark resolved.
- 01:05 PM Linux kernel client Bug #915: kclient corrupts on direct_io? ceph_direct_io_test reliably hoses the client machine
- Can't reproduce anymore, apparently some of the recent commits fixed this.
jobs 289 & 290 are both happy. Will re-ru... - 12:14 PM rgw Bug #965 (Resolved): rgw is crashing on ceph-playground
- appears to be fixed now
- 11:42 AM CephFS Feature #818 (Resolved): mds: robust lookuphash
- Ok, this passes my tests now, yay!
- 11:00 AM Bug #969 (Resolved): config: --name sdfsdf.sdfsdf should fail
- Passing a bad type to --name should cause it to give a usage message, rather than silently defaulting you to "client."
- 10:55 AM Bug #833 (Resolved): rbd wiki documentation uses old sysfs /sys/class interface
- 09:17 AM CephFS Tasks #922: fsstress: Request ping-pongs when dentry and inode auth are separate
- Sage's mds_ino search stuff made this pretty simple to deal with on my end!
- 04:32 AM Revision 5e4839b0 (ceph): makefile: include MMDSFindIno[Reply].h
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:10 AM Revision ddc06f11 (ceph): journaler: fix comments
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 01:09 AM Revision d5fa16a4 (ceph): Merge branch 'mds_ino'
- 01:07 AM Revision 30bffabd (ceph): mds: try find_ino_dir if find_ino_peers fails for lookuphash
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 01:07 AM Revision d31b76a6 (ceph): mds: find_ino_dir
- Search for a dir ino by checking the ondisk directory object path xattr.
Signed-off-by: Sage Weil <sage.weil@dreamho... - 12:12 AM Revision ce2fabc3 (ceph): common: fix cmdline of indashes_to_underscores
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:02 AM Revision 0a63567e (ceph): rgw: Properly support truncating on S3 PUT
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
03/31/2011
- 11:51 PM Revision 3f1e9b0f (ceph): mds: use helper to generate dir inode object names
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:42 PM Revision 26392894 (ceph): mds: allow explicit finisher context for path_traverse
- Previously we could only path_traverse and retry a request or message.
This just allows an explicit context to be use... - 11:38 PM Revision 34b0813d (ceph): mds: fix possible null dereference
- Fix null deref on forward and req (!mdr) caller.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 11:07 PM Revision 3afea8f1 (ceph): client: lookuphash
- Implement low-level lookuphash on client, for MDS testing.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 11:07 PM Revision c2c333d8 (ceph): mds: find_ino_peer
- Search for an ino on peer MDSs. Still need to handle mds failure/recovery
handling.
Signed-off-by: Sage Weil <sage.... - 11:07 PM Revision 39703405 (ceph): mds: try find_ino_peers for lookuphash requests
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:07 PM Revision 705161b1 (ceph): mds: add mds failure/recovery support for find_ino_peers
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:54 PM Revision cb2f966c (ceph): vstart.sh: don't put quotes in generated ceph.conf
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:31 PM Revision 07ffd55c (ceph): ceph_argparse: convert dashes to underscores
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:36 PM Revision 9567ba1a (ceph): man: Make </VirtualHost> be on its own line.
- Apparently, now there's a vertical paragraph space before it,
but that's better than being tacked at the end of the p... - 08:36 PM Revision e6544a37 (ceph): client: Use NULL instead of 0, for clarity.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 08:22 PM Revision 02581664 (ceph): rgw: fix buffer overflow
- Remove unecessary code that was causing a buffer overflow.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 06:44 PM Revision 4cc7dcc4 (ceph): mkcephfs: Actually do a mkfs.btrfs
- Signed-off-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Sage Weil <sage@newdream.net> - 06:39 PM Revision 96067588 (ceph): rgw: recover from partial bucket creates
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:33 PM Revision f895e20f (ceph): rgw: use tmap key (excl) create when creating a bucket
- This way we fail with EEXIST if the bucket already exists in the user
object.
Signed-off-by: Sage Weil <sage.weil@dr... - 06:15 PM CephFS Feature #818: mds: robust lookuphash
- commit:d5fa16a476b3248a89e1ef8956f5ab71390ec361
- 05:27 PM CephFS Bug #966 (Can't reproduce): mds: scatterstat error
- - vstart -d -n -x
- cfuse mnt
- cp /usr mnt/foo for a while (~15 mb of mds journal)
- stop the copy
- restart mds... - 04:47 PM Revision 7a2b6531 (ceph): config: remove kill_after
- It's unused.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 04:19 PM Tasks #916 (Resolved): fsstress results in failed CDentry assert
- 04:18 PM CephFS Tasks #923 (Resolved): Waiter does not get woken in fsstress
- 04:18 PM CephFS Tasks #928 (Resolved): Assert failure on replica: has auth pins in _logged_slave_rename
- 04:18 PM Tasks #921 (Resolved): Snaprealm issue
- 04:18 PM CephFS Tasks #934 (Resolved): auth_unpin assert fail in Locker::xlock_finish
- 11:47 AM rgw Bug #766 (Resolved): slow bucket creation times
- Okay, it now looks like:
- osd op to update user object bucket list
- osd op to create bucket object
- monitor o... - 11:41 AM Linux kernel client Bug #954 (Resolved): rbd: null pointer deref during osd_reset
- 10:48 AM rgw Bug #965 (Resolved): rgw is crashing on ceph-playground
- Bug'd as per TV's request. Not including stack trace here, since I can't be sure it doesn't include any information s...
- 10:43 AM Bug #964 (Resolved): librbd: snap rollback does not work
- 10:14 AM Bug #927 (Resolved): rbd image opened at a snapshot is writable
- 10:13 AM Bug #868 (Can't reproduce): librbd::AioCompletion::add_block_completion causes Qemu to crash
- 05:00 AM Bug #868: librbd::AioCompletion::add_block_completion causes Qemu to crash
- I've got my new cluster running now with all the latest code (Ceph and Qemu).
I've been doing some heavy writing i... - 09:44 AM Bug #955: remove max_open_files or make it do something
- Just to note, max_open_files is used in the startup scripts. kill_after is unused and has been removed.
- 12:55 AM Revision d5f10bcb (ceph): testlibrbd: fix warnings
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:54 AM Revision e44a5470 (ceph): testlibrbd: fix warnings
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:35 AM Revision 3fe957a5 (ceph): Merge branch 'next'
- 12:26 AM Revision 3946779b (ceph): librados: return -EROFS when trying to write to a snapshot
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:26 AM Revision ab6c0365 (ceph): librados: add selfmanaged_snap_rollback
- This was removed in 2cb86f713df38ebee6aa10a81157f99264a59a70, but is
required for selfmanaged snaps because their sna... - 12:26 AM Revision 7b28d5af (ceph): librados, librbd: add some debugging messages
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:26 AM Revision e424171c (ceph): librbd: fix snapshot handling
- To ensure consistency, always set the snap context when the header is
updated. If snapid is set, we update librados' ... - 12:26 AM Revision 5141e90b (ceph): testlibrbd: clean up and add tests for snapshots
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
03/30/2011
- 11:55 PM Revision 2b7cc153 (ceph): Merge branch 'next'
- Conflicts:
src/cmon.cc
src/osdc/Journaler.cc - 11:49 PM Revision 69f174c0 (ceph): mkcephfs: minor fix
- line 400: && $type == "osd" is not always valid.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 11:49 PM Revision cf3bb44d (ceph): mkcephfs: copy to daemon nodes for each daemon
- The tmp directory is removed after each daemon. Previously, this would
break if two daemons were on the same node. ... - 11:46 PM Revision 5a1ccdce (ceph): journaler: don't block when we adjust back write_pos
- is_readable() may need to adjust the write_pos backward, but will return
false. If we are at the end we still need t... - 11:36 PM Revision 887a01f6 (ceph): config: parse_argv fix
- md_config_t::parse_argv: fold md_config_t::parse_argv_part2 into
parse_argv. Fix brokenness introduced by the std::st... - 11:32 PM Revision c6f1cced (ceph): rados: print error message on exit
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:32 PM Revision ab8871ae (ceph): rados: add 'tmap {set,create} obj key value'
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:31 PM Revision 58fc0cf4 (ceph): osd: add tmap key (exclusive) create
- Return EEXIST if a key already exists.
Signed-off-by: Sage Weil <sage@newdream.net> - 10:44 PM Revision 5837a255 (ceph): config: fix use of uint64_t* as uint32_t*
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:41 PM Revision 6f27492e (ceph): mkcephfs: copy to daemon nodes for each daemon
- The tmp directory is removed after each daemon. Previously, this would
break if two daemons were on the same node. ... - 10:39 PM Revision 59923a8e (ceph): config: tighten up uint32_t use
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:04 PM Revision c3746dad (ceph): common_init: fix keyring loading
- Can't pass a NULL char* to std::string constructor.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 09:57 PM Revision a68ef601 (ceph): common: Add strict_strtof, strict_strtod, unittest
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:30 PM Revision 493e2d95 (ceph): OSD: ReplicatedPG: commits should be high priority, like acks.
- This was consistently breaking stuff for some people, as the acks were
high priority but the commits weren't. They sh... - 07:07 PM Revision 9cde1f44 (ceph): config: use std::string in md_config_t
- Use std::string to represent md_config_t strings. This makes memory
management a lot easier and should fix some leaks... - 05:08 PM Revision 3c94cc20 (ceph): test/bufferlist.cc: Added append tests
- Tests appending many randomly sized chunks many of which are size zero.
Signed-off-by: Samuel Just <samuel.just@drea... - 04:49 PM CephFS Tasks #934: auth_unpin assert fail in Locker::xlock_finish
- Should be good to go in commit:1b5e5d80f1278ae8156c912a7ad62e1661a13b7c
- 04:36 PM Bug #963: osd: errors part way through multiple ops leave inconsistent object context
- You're talking about ObjectOperations here? I vote for discard everything, since those are supposed to be more or les...
- 04:07 PM Bug #963 (Resolved): osd: errors part way through multiple ops leave inconsistent object context
- we bail out and do nothing, but leave behind a partial set of changes to the cached object context(s).
we need t... - 01:08 PM Bug #917 (Resolved): OSD sending out-of-order acks
- In commit:493e2d952ad24d8c8cab372e942ea3e18169ab4e.
It turns out that repop acks were getting CEPH_MSG_PRIO_HIGH, ... - 12:44 PM Bug #917: OSD sending out-of-order acks
- attaching osd4 logs from same start time until last osd_op...
- 12:37 PM Bug #917 (In Progress): OSD sending out-of-order acks
- Well, in this log the problem is that osd4 is acking client4168.0:8637 and then committing client4168.0:8636 (without...
- 01:05 PM Linux kernel client Feature #962 (Resolved): d_prune
- 01:00 PM Feature #961 (Resolved): obsync: -n mode
- show what you would do, but don't actually do it.
- 12:59 PM Bug #960 (Resolved): obsync: support rados pool "buckets"
- - sync directly to/from librados
- copy amazon acl's into same xattr name that radosgw uses, so that this is equival... - 12:32 PM Bug #959 (Resolved): use sizeof rather than if (0) in logging macros (and other macros)
- Perhaps we need to use sizeof to prevent gcc from generating code that should be if (0)'ed out?
Not sure, but it c... - 12:25 PM Bug #942 (Resolved): eliminate NULL/empty string confusion in g_conf
- implemented by commit:9cde1f44a57cd630a66e77ed3d59127d91834d09
- 09:24 AM Feature #958 (Resolved): rbd: map/unmap kernel rbd device via rbd(?) tool
- rbd devices are mapped/removed via /sys/bus/rbd. This is a bit tricky, especially if you need to set up the security...
- 08:48 AM Linux kernel client Bug #954 (In Progress): rbd: null pointer deref during osd_reset
- 08:20 AM Revision 172835ff (ceph): ConfUtils: move parsing into config.cc
- Move parsing into config.cc, since there was already parsing code there.
Move metavariable escaping out of ConfUtils;... - 12:04 AM Revision 1fb543cb (ceph): ConfUtils: fix up read API a bit
- Remove def_val. Return error codes for not found, parse error, etc.
Fix API users.
Signed-off-by: Colin McCabe <coli... - 12:04 AM Revision a854feb0 (ceph): ConfUtils: decode to std::string
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
03/29/2011
- 11:31 PM Revision 44100001 (ceph): mount.ceph: Use LDADD not LDFLAGS to bring in -lkeyutils.
- Without this, flab (gcc 4.2.4) failed with:
libtool: link: gcc -Wall -D__CEPH__ -D_FILE_OFFSET_BITS=64 -D_REENTRANT ... - 11:23 PM Revision a2ec936a (ceph): cosd: fix bind to specific address
- Do not default to public_addr for cluster, heartbeat messengers.
Signed-off-by: Sage Weil <sage@newdream.net> - 11:16 PM Revision 3a1a3185 (ceph): mkcephfs: minor fix
- line 400: && $type == "osd" is not always valid.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 10:37 PM Revision 92e0751f (ceph): Add an RPM BuildRequires on keyutils-libs-devel.
- See 9aab8c6b051469904d7ecd5a9a6dd7aaf5cc3931.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 09:53 PM Revision 4af786c2 (ceph): config: temporary fix for monclient conf issue
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:49 PM Revision 138f360d (ceph): cmon: fix typo
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:47 PM Revision 1601df65 (ceph): Merge branch 'stable' into next
- Conflicts:
configure.ac
debian/changelog
src/msg/SimpleMessenger.cc - 09:11 PM Revision 0f9e318b (ceph): mount.ceph: fix missing includes
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:34 PM Revision 9aab8c6b (ceph): Add a Debian Build-Dep on libkeyutils-dev.
- This is needed by bee85518e2885cc93fe8ca634292ad4846515456.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.... - 07:55 PM Revision 73671c84 (ceph): common: Make armor.h safe to use from C.
- mount.ceph needs to base64-decode the secrets, so we can get rid of
the kernel-side base64 decode, but it doesn't nee... - 07:55 PM Revision bee85518 (ceph): mount.ceph: Use kernel key management API when possible.
- Backwards compatible with older kenrnels, for now.
Fixes: #852
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamho... - 07:55 PM Revision db3703ac (ceph): mount.ceph: Fix strncmp length comparing "secretfile".
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 07:55 PM Revision 9233889a (ceph): mount.ceph: Silently pass through options "name" and "secret".
- Fixes: #914
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> - 07:55 PM Revision 69f24671 (ceph): mount.ceph: Modprobe ceph before trying the mount.
- This will be needed for the next few commits, where we try to load the
keys into the kernel; without ceph.ko loaded, ... - 07:39 PM Revision ab03eda6 (ceph): rgw: add bucket to user list prior to creating bucket
- This avoids hitting the osd immediately following a bucket creation, which
will avoid any map update latency for most... - 07:11 PM Revision e8ceadb8 (ceph): objecter: add create to ObjectOperation, known strings
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:11 PM Revision 9d0697a3 (ceph): librados: add create to ObjectOperation
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:11 PM Revision 50ad51a5 (ceph): rgw: create bucket object and xattr in single operation
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 07:08 PM Revision 12a82a5a (ceph): SyntheticClient: const cleanup
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:05 PM Revision 59e2ef1e (ceph): FileStore: convert dev, jdev to std::string
- Convert dev, jdev to std::string in preparation for adding std::string
to g_conf.
Signed-off-by: Colin McCabe <colin... - 06:58 PM Revision 41899a3b (ceph): cmon: add --inject-monmap option
- This lets you manually inject a monmap into a down monitor. This is useful
in cases where you need to change the mon... - 06:20 PM Revision de6338c4 (ceph): mkcephfs: don't clobber osdmap when importing crushmap
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:14 PM Revision dd5063bd (ceph): librados: compound ObjectOperation
- Wrap the internal ObjectOperation and expose via librados.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:32 PM Bug #955 (Won't Fix): remove max_open_files or make it do something
- 09:39 AM Bug #955: remove max_open_files or make it do something
- Likewise with kill_after. Perhaps this was left in after some debugging?
- 09:38 AM Bug #955 (Won't Fix): remove max_open_files or make it do something
- md_config_t::max_open_files seems to be unused by anyone?
- 04:16 PM Bug #957 (Resolved): Binding to specific ports is not supported, yet config & code tries to do so
- commit:a2ec936a7cd1c3f98b5da1769eb88a280f40404a
- 03:13 PM Bug #957 (Resolved): Binding to specific ports is not supported, yet config & code tries to do so
- Having ceph.conf say
[osd.0]
public addr = 10.1.2.3:4567
Makes cosd try to bind multiple listening sockets to... - 02:34 PM Linux kernel client Bug #954 (Resolved): rbd: null pointer deref during osd_reset
- 09:08 AM Linux kernel client Bug #954 (Resolved): rbd: null pointer deref during osd_reset
- ...
- 12:59 PM Linux kernel client Bug #852 (Resolved): secret key shows up in /etc/mtab after mount -o secretfile=/.../key
- commit bee85518e2885cc93fe8ca634292ad4846515456
Author: Tommi Virtanen <tommi.virtanen@dreamhost.com>
Date: 2011-... - 12:59 PM Bug #914 (Resolved): mount.ceph should not complain about options name, secret
- commit 9233889a7be8f672729a71e48681a1e779dabd1c
Author: Tommi Virtanen <tommi.virtanen@dreamhost.com>
Date: 2011-... - 12:17 PM Feature #945 (Resolved): librados: mutate api
- commit:9d0697a36e9c7381943d8ca2e68dae73f89c9a53
- 11:17 AM Bug #956: obsync: preserve ACLs when doing bucket to bucket transfers
- - this should be an option, probably -p (and implied by -a) to be consistent with rsync and cp
- for files, yea, we ... - 11:06 AM Bug #956 (Resolved): obsync: preserve ACLs when doing bucket to bucket transfers
- Should we preserve ACLs when doing bucket to bucket transfers? Probably.
I'm not sure how we should handle ACLs wh... - 10:19 AM Bug #898 (Resolved): vstart.sh fails if host has many IP addresses
- Fixed in commit fc81de5fc84a8afdedd4510f95168fc8e0716a0a.
- 09:53 AM rgw Bug #766: slow bucket creation times
- Two parts here. The first is the bucket creation, and goes like:
- bucket object .. create? class op?
- bucke...
03/28/2011
- 11:52 PM Revision eba54376 (ceph): config: refactor conf_set_conf_val a bit
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:28 PM Revision 82039d7d (ceph): config: be stricter in set_val
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:23 PM Revision f18eac09 (ceph): config: whitespace, code ordering cleanup
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:14 PM Revision 51e21a71 (ceph): common: more argument parsing into ceph_argparse
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:14 PM Revision ddcd6415 (ceph): common: rename old argument parsing macros
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:50 PM Revision fc81de5f (ceph): vstart.sh: Filter out IPv6 and localhost IP addresses.
- On e.g. Ubuntu 10.10, hostname --ip-address outputs something
like "::1 10.1.2.3 127.0.1.1", and this makes the gener... - 10:48 PM Revision d8096e75 (ceph): common: separate file for strict_strtol
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:14 PM Revision 5c218237 (ceph): objecter: ObjectOperation fixes
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:03 PM Revision 98dd2d11 (ceph): journaler: remove obsolete journaler_cache thing
- This is ancient code used for testing impact of journal IO, no longer
useful.
Signed-off-by: Sage Weil <sage.weil@dr... - 05:23 PM Bug #917: OSD sending out-of-order acks
- I think this catches it... has osd_op's in it and they should stop at the crash.
Log excerpt starts 20min before cra... - 03:44 PM Bug #953 (Resolved): osd, mon: use sys_syncfs() when it hits glibc
- 02:27 PM Bug #952: Make lack of btrfs/user_xattr more obvious in mkcephfs
- We should just be able to switch this from dout to derr, correct?
- 02:15 PM Bug #952 (Resolved): Make lack of btrfs/user_xattr more obvious in mkcephfs
- Currently, cosd has this on startup:...
- 02:06 PM Tasks #888 (Resolved): get new sepia machines into autotest pool
- 02:04 PM Feature #951 (Resolved): set up semi-permanent cosd cluster, dev backups
- - fs on cosd machines
- one permanently allocated (sepia?) machine that mounts it
and then
- nightly cron jobs to ... - 11:58 AM Bug #946 (Resolved): RPM spec missing a few installed files
- 11:16 AM Bug #946 (Resolved): RPM spec missing a few installed files
- In trying to build an RPM for 0.25.2 on RHEL5 x86_64, it gets most of the way done, then fails because it installs a ...
- 11:51 AM Bug #822 (Resolved): fix i386 build of radostestpp
- 11:51 AM Bug #816 (Can't reproduce): fs size underflowed
- 11:50 AM rgw Feature #786 (Rejected): journal rgw operations to backup storage
- 11:50 AM Bug #744 (Won't Fix): osd class changes take too long to propagate
- 11:50 AM Bug #674 (Can't reproduce): tiobench stress test , OSD timeout
- 11:48 AM Bug #577 (Resolved): unify PG creation code in OSD::handle_pg_notify and OSD::_process_PG_info
- 11:48 AM rgw Feature #9 (Won't Fix): Access unimported data
- 11:48 AM rgw Bug #353 (Resolved): Handle non-ascii filenames
- 11:44 AM Bug #145 (Resolved): Check build dependencies for FastCGI
- 11:29 AM Linux kernel client Feature #949 (Rejected): rbd: async writes, flush/barrier
- 11:25 AM Feature #948 (Rejected): autotest: graph rbd performance
- iozone?
- 11:21 AM Feature #947 (Resolved): librbd: async writes, honor flush
- see thread on qemu-devel list from last year
- 11:13 AM Feature #901 (Resolved): autotest: get all 4 racks of sepia online
- 11:05 AM Messengers Bug #882 (In Progress): misc leaks in librados
- 11:01 AM Bug #924 (Can't reproduce): Various OSD/FileStore bugs
- 10:24 AM Bug #799 (Closed): mkcephfs error
- 10:23 AM CephFS Cleanup #886 (Resolved): journaler: rip out ack_barrier cruft
- 10:21 AM CephFS Cleanup #929 (Resolved): mds: make journal replay readahead less silly
- 10:16 AM Feature #945 (Resolved): librados: mutate api
- support compound operations in librados, e.g. "write this extent, set this xattr"
- 09:45 AM rbd Bug #944 (Duplicate): rbd snapshots not exposed as read-only
- I'm not sure what level the issue exists at, but attaching an RBD snapshot using virtio makes it look read-write on a...
03/26/2011
03/25/2011
- 11:02 PM Bug #943: 3-mon cluster won't start
- Oh, and please, no mocking of my slow 100Mbps home network ;-)
I'm using ceph for data replication, not for speed.... - 10:50 PM Bug #943: 3-mon cluster won't start
- Confirmation that the problem has to do with the length of logm messages (or rather the delay in transmitting them). ...
- 08:51 PM Bug #943: 3-mon cluster won't start
- Oops, I just realized the title is a bit too general. won't always restart from scratch would probably be more accur...
- 08:49 PM Bug #943 (Resolved): 3-mon cluster won't start
- I had run into this before, when I had only two monitors up, and thought the monitors had gone out of sync due to btr...
- 10:39 PM Revision 6966c3ed (ceph): msgr: move pidfile handling, signal handing out
- Move signal handling and pidfile writing out of the messenger and into
signal.cc and pidfile.cc, respectively.
(Try ... - 09:45 PM Revision b4d49df4 (ceph): testradospp: zero terminate before printing strs
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:34 PM Revision e1ce7675 (ceph): mds: include .ceph is root directory
- If the dentry isn't marked dirty _commit_partial won't save it. This is
caught later by the check_rstats() (or anyon... - 09:04 PM Revision 0f241fcd (ceph): common: run_cmd: fix bad implicit conversion
- Since NULL is really just a macro defined to be 0, we must use
(char*)NULL or similar to force the compiler to use a ... - 08:50 PM Revision 009714b8 (ceph): mds: fix client session removal on journal replay
- We want to remove the client session from the map as long as it is not
attached to an actual messenger Connection. T... - 07:58 PM Revision 45cc761c (ceph): Merge remote branch 'vapre/journaler'
- 07:54 PM Revision 5658857a (ceph): Fix Debian librados1 -> librados2 package renames
- The librados2{,-dbg,-dev} packages should conflict with and replace their
librados1{,-dbg,-dev} counterpart. - 07:37 PM Revision ba4678da (ceph): journaler: remove ack/safe distinction
- Rip out old complexity to _only_ pay attention to when data is safely
committed on disk. No more ack/safe distinctio... - 07:30 PM Revision 2a53c83c (ceph): obsync: Remove deadcode
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:14 PM Revision e8d86432 (ceph): rgw: if formatter gets too much data, print some
- If the RGW formatter gets too much data, print some data rather than
printing nothing. This is safe because vsnprintf... - 07:12 PM Revision a5be3726 (ceph): rgw: implement JSON escaping
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:21 PM Bug #917: OSD sending out-of-order acks
- I'm attaching a cfuse log and an osd log.
Is the transaction id supposed to be showing up in the osd?
From inspecti... - 04:00 PM Bug #917: OSD sending out-of-order acks
- Updating bug name/description to match actual problem.
- 01:49 PM Bug #917: OSD sending out-of-order acks
- Apologies... didn't track all the way back to the correct transaction.
- 11:50 AM Bug #917: OSD sending out-of-order acks
- Here's the log from tid (osd tid) 24101 and forward.
Thanks for your assistance.
- 05:46 PM Revision 1bfabefe (ceph): rgw: init_rest: small refactoring
- The validation function has been updated to handle empty bucket names.
Empty key names also pass validation.
Signed-... - 05:37 PM Revision ede7e3df (ceph): units: remove "performance test" from crypto units
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:37 PM Revision 50d0947a (ceph): units: fix cauthtool usage test
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:33 PM Revision 1e7b371c (ceph): rgw: escape XML attributes
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:23 PM Revision 7eca830b (ceph): boto_tool.py: usage fix
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:23 PM Revision aac1ef57 (ceph): rgw: don't reject requests that don't have bucket
- Some requests don't specify a bucket. Allow those!
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 04:52 PM Revision e1457a28 (ceph): journaler: remove some dead code from read_entry()
- These were used to track completion.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:51 PM Revision a0704eb3 (ceph): journaler: issue separate reads per period
- This lets us potentially digest any read data as soon as possible. Before
the Filer would issue a string of reads and... - 04:42 PM Revision c9d0edeb (ceph): journaler: fix prefetch to handle multiple in-flight reads
- If we issue multiple reads, they can come back in any order.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:30 PM Revision 31293194 (ceph): journler: make readahead/prefetch smarter
- Always try to prefetch N segments ahead of the current read position. The
old implementation would read a bunch of d... - 04:29 PM Revision 2ecbd024 (ceph): journaler: reduce prefetch interval
- 50MB was overkill.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:25 PM CephFS Tasks #934 (In Progress): auth_unpin assert fail in Locker::xlock_finish
- Turning up other issues as a result of the cleanup. (Apart from the obvious one where hacking the auth_pins directly ...
- 03:57 PM CephFS Tasks #934: auth_unpin assert fail in Locker::xlock_finish
- Hmm, so apparently it actually is auth when we finish, but it wasn't auth when we started. commit:f46faeed51c6547600d...
- 03:56 PM Bug #924: Various OSD/FileStore bugs
- os/FileStore.cc: In function 'unsigned int FileStore::_do_transaction(ObjectStore::Transaction&)', in thread '0x7fab3...
- 03:44 PM Bug #942 (Resolved): eliminate NULL/empty string confusion in g_conf
- Right now there are two ways to indicate that a string has "no value" in g_conf: either you can set a string's value ...
- 03:39 PM Revision 921cb061 (ceph): journaler: remove unused read_entry()
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 02:28 PM CephFS Bug #938 (Resolved): mds: rstats mismatch
- commit:e1ce76750b016b1c3af50497d1bb98663829504d
- 09:40 AM CephFS Bug #938 (Resolved): mds: rstats mismatch
- easily triggering this with ./csyn --syn makedirs 4 4 4 and restarting mds after a minute or so...
- 12:21 PM Feature #883 (Resolved): osync tool
- implemented, yay!
- 12:21 PM Bug #941 (Resolved): obsync: escape object names that end in slash
- when syncing with a local filesystem, osync should change object names that end in slash to end in something else. Ma...
- 12:18 PM rgw Feature #940 (Closed): obsync: use COPY optimization
- The COPY operation can copy data between buckets without involving the client.
In Amazon's words,... - 12:16 PM Bug #908 (Resolved): RGW allows users to create buckets and objects with invalid names
- Implemented all subtasks.
- 12:16 PM Tasks #939 (Resolved): properly escape JSON in RGW
- implemented by commit:e8d86432d47316c681a30347ae43d011af6d8ea3
- 10:39 AM Tasks #939 (Resolved): properly escape JSON in RGW
- properly escape JSON
Here we go again... - 11:27 AM Bug #615 (Resolved): osd: improve op+journal throttling
- 10:39 AM Bug #931: opening up security. xattr
- I'd like to implement the same code from the kernel client as mentioned above.
However, to use those macro definitio... - 10:35 AM Tasks #919 (Resolved): properly escape XML
- implemented by commit:1e7b371c28009e9f5e8e1126ff71f7575dd13c40
- 10:11 AM CephFS Bug #937 (Closed): journaler: _issue_read assert
- fixed
- 08:44 AM CephFS Bug #937 (Closed): journaler: _issue_read assert
- ...
- 12:28 AM Revision 658b1404 (ceph): common_init: allow init without config file
- Add a flag, CINIT_FLAG_NO_DEFAULT_CONFIG_FILE, that specifies that the
program should not read a config file by defau...
Also available in: Atom