Project

General

Profile

Activity

From 03/04/2012 to 04/02/2012

04/02/2012

08:24 PM Revision addc7446 (ceph): rgw: check for subuser existence
This fixes #1856: looking up subuser that doesn't exist returns
user as long as subuser prefix defined existing user....
Yehuda Sadeh
02:30 PM rgw Bug #1853 (Resolved): rgw: qa test to verify bucket recreation does not override bucket
Implemented, commit:1551c5b08714b415c49fc759002b7c6a6d4d611a. Yehuda Sadeh
01:26 PM rgw Bug #1856 (Resolved): It is possible to look up an rgw user by a subuser that does not exist as l...
Fixed, commit:addc744692f60885a747c4531cd12bf19b3a7f2a. Yehuda Sadeh
11:15 AM rgw Feature #2171: rgw: asynchronously calculate md5
Thinking about it some more, it's probably not the best use of time and effort. We initiate the md5 calculation after... Yehuda Sadeh
08:29 AM Bug #2178: rbd: corruption of first block
Hi Sage,
here we go again, with ceph-0.44.1-1-g41f84fa
One bad file with following infos:
20120402 171642.12...
Oliver Francke
12:04 AM Revision e792cd93 (ceph): filestore: fix ZERO fallback write
It helps if we write zeros!
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil

04/01/2012

11:24 PM Revision 8434caf5 (ceph): qa: test_rewrite_latency
Tool to measure latency of overwriting a single block.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
07:23 PM Bug #2221: Monitor setup bugs
2) ... Greg Farnum
06:35 PM rbd Feature #2232: qemu: resize guest disk when rbd image is resized
I tested this on Friday, and qemu rereads the size (at least when using virtio) when the guest requests it (i.e. echo... Josh Durgin
04:21 PM rbd Feature #2232 (New): qemu: resize guest disk when rbd image is resized
According to Christoph, this is probably just a matter of calling bdrv_truncate() with the new size. If that doesn't... Sage Weil
04:19 PM rbd Feature #2231 (Resolved): librbd: expose header change (resize?) via api
we need a callback or something so that users (qemu) can be informed when the header changes. this will let them, sa... Sage Weil

03/31/2012

03:22 PM Feature #1655: gitbuilder aggregator page
I took some inspiration from the updated aggregator script that is now at http://ceph.newdream.net/gitbuilder.cgi. I'... Jimmy Tang
03:31 AM Revision dbc70b9d (ceph): Merge remote branch 'gh/wip-mon_setup'
Reviewed-by: Sage Weil <sage@newdream.net> Sage Weil
03:18 AM Revision f8a53869 (ceph): osd: fix error code return from class methods
Don't shadow the result at function scope.
Fixes: #2148
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
12:22 AM Revision 15f0a327 (ceph): monmaptool: make clear you can set the fsid when making a new map.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com> Greg Farnum
12:07 AM Revision 208daeb3 (ceph): ceph_mon: fix fsid parsing.
fsid is a field in the CephContext _conf structure and is parsed by
the standard options parsing library before it ge...
Greg Farnum

03/30/2012

11:15 PM Revision 9a69c3f3 (ceph): ceph.conf: enable 'osd recover clone overlap'
to test the recovery cloning in qa. this was redone, but forgot to enable
it in qa.
Sage Weil
11:14 PM Revision aa31035e (ceph): osd: update_stats() on reads too
Update pg stats on any op completion (read or write), not just writes. Do
the calls with log_op_stats() for consiste...
Sage Weil
11:11 PM Revision 28788654 (ceph): log: dump_recent in fatal signal handler
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
11:07 PM Revision f27acbc8 (ceph): Merge remote-tracking branch 'gh/wip-log'
Conflicts:
src/common/config_opts.h
Sage Weil
11:00 PM Revision 374bef9c (ceph): Merge remote branch 'gh/wip-osd-hb'
Sage Weil
10:37 PM Revision f7f65ebe (ceph): osd: fix typo in debug message
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
09:57 PM Revision 75e3b9b3 (ceph): Merge remote branch 'gh/wip-osd-recovery-sources'
Sage Weil
09:23 PM Revision df5860fe (ceph): objectstore: name Sequencers
Assign a (unique) name to each Sequencer. This will aid in debugging, and
can be useful when dumping traces of FileS...
Sage Weil
09:11 PM Cleanup #2230 (Resolved): deprecate 'btrfs devs'
Sage Weil
09:00 PM rgw Feature #2229 (New): rgw: functional tests for rgw class
A series of simple functional tests to verify the rgw class methods behave as they should. Sage Weil
08:58 PM Bug #2148 (Resolved): osd: class error return not propagated to client
commit:f8a53869f6db4c76516ee525f00f87f930920692 Sage Weil
06:57 PM Revision 29c01f25 (ceph): ceph_common.sh: Remove dead code.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> Tommi Virtanen
06:27 PM Revision ba6bb4cf (ceph): man: Oops, update ceph-mon(8) for real. Sorry about that.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> Tommi Virtanen
06:26 PM Revision 541a543c (ceph): man: Update ceph-mon(8) after reStructuredText syntax fixes.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> Tommi Virtanen
06:16 PM Revision 2c542442 (ceph): doc: Remove duplicate anchor from (unused) overview doc.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> Tommi Virtanen
06:11 PM Revision 1ec47db1 (ceph): doc: Convert the mailing list mention to not be a section heading.
If toctree is inside a section, the subtree is inside the section too.
We don't want all of dev/* to be under "Mailin...
Tommi Virtanen
06:11 PM Revision b162696b (ceph): doc: Fix reStructuredText syntax errors.
Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com> Tommi Virtanen
05:52 PM Revision 2d1a96d3 (ceph): add include/stringify.h
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
05:27 PM Bug #2221: Monitor setup bugs
(1) is a problem due to options parsing collisions...fixed!
(2) is directly contradicted by my testing...?
(3) I ne...
Greg Farnum
04:59 PM Revision b25817a5 (ceph): FileJournal: check pwrite return value when zeroing journal
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
04:52 PM Revision 41f84fac (ceph): filestore: set guard on collection_move
During recovery we submit transactions like:
- delete a/foo
- move tmp/foo to a/foo
This prevents the EEXIST chec...
Sage Weil
04:25 PM Bug #2026 (Can't reproduce): osd: ceph::HeartbeatMap::check_touch_file
Sage Weil
04:25 PM Bug #2045 (Can't reproduce): osd: dout_lock deadlock
haven't seen this in a while.
also, this code is about to go away anyway with wip-log.
Sage Weil
04:16 PM Bug #2102 (Can't reproduce): osd: pg stuck in backfill
Sage Weil
04:15 PM Bug #2102 (Duplicate): osd: pg stuck in backfill
Sage Weil
04:14 PM Bug #2002: osd: racy push/pull for clones
i take that back; this wasn't enabled in qa. adding to the teuthology ceph.conf file. Sage Weil
04:12 PM Bug #2002 (Resolved): osd: racy push/pull for clones
haven't seen this in forever; looks fixed. Sage Weil
04:11 PM Bug #2209 (Resolved): osd: read kb stats not tracked?
commit:aa31035e555129e56888320b84f16264f28bd7df Sage Weil
03:59 PM Bug #2116 (Resolved): Repeated messages of "heartbeat_check: no heartbeat from"
fixed by commit:374bef9c97266600b4c6b83100485d7250363213 Sage Weil
03:59 PM Bug #2165 (Resolved): osd: recovering ending with missing
fixed with merge of commit:75e3b9b309e5365975e3e5855c065bd4fe28b64c Sage Weil
03:58 PM Bug #2178: rbd: corruption of first block
Sage Weil
02:51 PM Bug #2178: rbd: corruption of first block
Please build the current git stable branch, which includes 41f84fac1ae4b4c72bf9bfe07614c4066c916fd1. The version sho... Sage Weil
07:35 AM Bug #2178: rbd: corruption of first block
Here the remaining timestamps from the other VM's with bad blocks:
VM-2:
20120330 105139.579830 filling block 171...
Oliver Francke
07:12 AM Bug #2178: rbd: corruption of first block
Hi *,
I needed a couple of runs, but managed now to provide some 81MiB/97MiB osd.X.log-files, where in between sh....
Oliver Francke
03:58 PM Bug #2164 (Resolved): osd: scrub missing _, snapset attrs
commit:41f84fac1ae4b4c72bf9bfe07614c4066c916fd1 Sage Weil
03:49 PM Revision f89f98df (ceph): osd: clear RECOVERING on start_peering_interval
This prevents us from, say, getting into a recovering+stray state.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
03:45 PM Revision 3cdd8d58 (ceph): osd: more heartbeat debug
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
03:45 PM Revision e1a58912 (ceph): osd: discard heartbeat_peer in note_down_osd
Discard the heartbeat_peer as soon as we find out, along with queued
failures, or else the heartbeat_check may come a...
Sage Weil
03:45 PM Revision 21e6e2b8 (ceph): osd: ignore peer epoch of 0 on ping reply
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
03:45 PM Revision efc27f19 (ceph): osd: don't fail new heartbeat peers
last_tx may be 0 because we just added this peer; don't mark them down
yet!
Signed-off-by: Sage Weil <sage.weil@drea...
Sage Weil
03:45 PM Revision 33b9187a (ceph): osd: rename hbin -> hbclient, hbout -> hbserver
This is way less confusing.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
03:44 PM Revision 4e2f0d14 (ceph): osd: simplify heartbeat logic
Simplify heartbeats to use a simple request/reply model.
- avoid any weirdness with map update timing
- no from/to...
Sage Weil
03:44 PM Revision fe5f0331 (ceph): osd: send pings from hbin
Fixes: #2212
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
02:32 PM Revision eebc9ec2 (ceph): test: test_workload_gen: Add callback for collection destruction.
When we remove a collection, we must cleanup after the coll_entry_t we
once had on the available collections set. For...
Joao Eduardo Luis
01:53 PM Revision 424b5b07 (ceph): ceph: --concise by default, add --verbose option
It's time.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
12:50 PM Feature #2227 (Closed): QA: create a test to verify operation with non-default layouts
I submitted a patch that modified ceph_calc_file_object_mapping()
in the ceph client, and when reviewing it Sage poi...
Alex Elder
09:53 AM Feature #2226 (Resolved): osd: better filestore idempotency test
... Sage Weil
03:31 AM Revision 409b648b (ceph): config: drop old debug_* items
...and replace code references with conf->subsys.should_gather().
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
02:16 AM Revision 5d981b15 (ceph): rgw: add unittest just to verify we link
This will flush out references to stuff in libglobal.la, among other
things.
Signed-off-by: Sage Weil <sage.weil@dre...
Sage Weil
02:06 AM Revision 69b01726 (ceph): config: fix librados, libcephfs unit tests
No more g_conf->debug.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
01:08 AM Revision 394d8b1e (ceph): Add test for object source marked down
Samuel Just
01:08 AM Revision b4aa098f (ceph): make Thrasher not inherit from Greenlet
Samuel Just
01:02 AM Revision 1c8ec702 (ceph): PG,ReplicatedPG: update missing_loc_sources with missing_loc
In some cases missing_loc was updated without missing_loc_sources
Signed-off-by: Samuel Just <samuel.just@dreamhost....
Samuel Just
01:02 AM Revision 05ef3ba6 (ceph): ReplicatedPG: fix loop in check_recovery_sources
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
12:35 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
I think I can be optimistic :)... Wido den Hollander

03/29/2012

10:06 PM Bug #2178: rbd: corruption of first block
Okay, I suspect this is actually bug #2164, which was causing the _ xattr to get lost when ceph-osd restarts on non-b... Sage Weil
09:52 PM Bug #2225 (Resolved): gitbuilder.ceph.com returning 503: Service Temporarily Unavailable.
I can't find any 503 in the apache logs on this machine. Could it be on the client side? Sage Weil
09:48 PM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
Well, I fixed one problem, but I can't see how it could have resulted in the log you posted.
Pushed a few more pat...
Sage Weil
11:36 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
I collected logs from 4 OSDs, they can be downloaded at: http://logger.ceph.widodh.nl/ceph/issues/2212/
At 10:13 t...
Wido den Hollander
09:21 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
Der.. do you have a log you can attach/post? Sage Weil
02:59 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
I reverted the extra debugging for the heartbeat stuff, but that didn't seem to consume all the CPU time.
The load...
Wido den Hollander
01:40 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
I just installted the code on my cluster and things do not seem to behave yet.
The cluster is still jumping around...
Wido den Hollander
08:54 PM Linux kernel client Bug #1940 (Resolved): locking cycle in ceph_osdc_start_request
commit:ab434b60ab07f8c44246b6fb0cddee436687a09a Sage Weil
08:15 PM Revision 41a09bea (ceph): Merge remote branch 'upstream/wip_latency'
Samuel Just
07:53 PM Linux kernel client Bug #1793 (Can't reproduce): NULL pointer dereference at try_write+0x627/0x1060
Marking this Can't Reproduce. Will reopen if it shows up again. Alex Elder
03:21 PM Linux kernel client Bug #1793: NULL pointer dereference at try_write+0x627/0x1060
Another 100 iterations of kernel_untar_build.sh using the current
master branch (c666601a935b94cc0f3310339411b6940de...
Alex Elder
07:51 AM Linux kernel client Bug #1793: NULL pointer dereference at try_write+0x627/0x1060
Bugs 1793 and 2081 have a signature of a page fault/bad memory reference
from process_one_work() -> con_work(), and ...
Alex Elder
07:53 PM Linux kernel client Bug #2069 (Can't reproduce): client crash during kernel_untar_build rm -r step
I just finished at least 150 iterations of kernel_untar.sh and never
hit this using the current master branch of cep...
Alex Elder
07:51 PM Linux kernel client Bug #2081 (Can't reproduce): msgr: spinlock badness?
Marking this Can't Reproduce. Will reopen if it happens again. Alex Elder
07:43 PM Linux kernel client Bug #2081: msgr: spinlock badness?
Another 100 iterations of kernel_untar_build.sh using the current
master branch (c666601a935b94cc0f3310339411b6940de...
Alex Elder
07:51 AM Linux kernel client Bug #2081 (Need More Info): msgr: spinlock badness?
Bugs 1793 and 2081 have a signature of a page fault/bad memory reference
from process_one_work() -> con_work(), and ...
Alex Elder
07:50 PM Linux kernel client Bug #2174 (Can't reproduce): rbd: iozone thrashing failure
OK, I'll go ahead and state that I can't reproduce this... Alex Elder
07:46 PM Linux kernel client Bug #2174: rbd: iozone thrashing failure
Status was Verified. Changing it to Need More Info because I can't even
seem to reproduce it at this point. (I sup...
Alex Elder
07:44 PM Linux kernel client Bug #2174: rbd: iozone thrashing failure
Another 12 iterations of suites/iozone.sh using the current
master branch (c666601a935b94cc0f3310339411b6940de751ba)...
Alex Elder
07:59 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
I don't know whether we've adequately captured the signature or symptoms
of this problem. I believe though that it ...
Alex Elder
07:20 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
I have been trying to reproduce this using the latest testing/master/for-linus
branch (they're the same right now) a...
Alex Elder
02:34 PM Revision c39ed568 (ceph): test: test_workload_gen: Fixing a memleak.
Apparently, the FileStore does not cleanup after transactions once they
are applied, which may lead to huge memory le...
Joao Eduardo Luis
09:27 AM Linux kernel client Bug #2224 (Rejected): Oops in __cfh_to_dentry
I setup an HA pair of NFS servers which re-export Ceph to NFS clients.
The HA pair is in active/standby mode, using...
Henry Chang
07:42 AM Feature #2087: lightweight filestore workload generator
Memory leak fixed.
Apparently, the FileStore does not cleanup after transactions once they are applied, which may ...
Joao Eduardo Luis
06:21 AM Feature #2087 (In Progress): lightweight filestore workload generator
Looks like some memory should be leaking bad, such that valgrind hangs on exit.
==19080==
==19080== HEAP SUMMARY...
Joao Eduardo Luis
07:24 AM Linux kernel client Bug #2064 (Resolved): ceph-client: messenger: nocrc flag not implemented correctly
Linus pulled in the changes without any immediate trouble, so
I'm marking this and a few others resolved.
Alex Elder
07:12 AM Linux kernel client Bug #2157 (Resolved): ceph: xattr: fix nanosecond display on i_rctime
Linus pulled in the changes without any immediate trouble, so
I'm marking this and a few others resolved.
Alex Elder
07:12 AM Linux kernel client Bug #2156 (Resolved): ceph: xattr: fix a possible buffer overrun bug
Linus pulled in the changes without any immediate trouble, so
I'm marking this and a few others resolved.
Alex Elder
07:11 AM Linux kernel client Bug #2155 (Resolved): ceph: xattr: wrong value assumed for "no preferred PG"
Linus pulled in the changes without any immediate trouble, so
I'm marking this and a few others resolved.
Alex Elder
05:56 AM Feature #2223 (Resolved): Tracing facility on FileStore
Allow a user to specify a file onto which log the transactions that come through OSDs' FileStores.
This should all...
Joao Eduardo Luis
05:47 AM Revision b3069e50 (ceph): ceph_argparse: drop useless declaration from unit test
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:16 AM Revision 4269f8d5 (ceph): ReplicatedPG: ctx might not contain an OpRequest
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
12:16 AM Revision 135a11ba (ceph): FileJournal: optionally zero journal on create
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
12:15 AM Revision 2486c61a (ceph): FileStore: Pass OpRequestRef into filestore in queue_transaction
This allow us to track op progress through the filestore.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Samuel Just
12:15 AM Revision d026cdc7 (ceph): FileJournal: use DSYNC for directio path
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
12:15 AM Revision 533bbf7b (ceph): osd/: OpRequest implements TrackedOp for passing into filestore
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just

03/28/2012

11:12 PM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
Ah, I see the bug now. Pushed a fix to wip-osd-hb, thanks!
Let us know if this behaves for you.. if so I'll pull ...
Sage Weil
04:23 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
It's quite large (222MB), so I uploaded the file, available at: http://logger.ceph.widodh.nl/ceph/osd.1.log_27-03-201... Wido den Hollander
10:51 PM Bug #2165: osd: recovering ending with missing
see wip-osd-recovery-sources Sage Weil
10:46 PM CephFS Bug #1811: 2 pjd chown tests failed on cfuse
... Sage Weil
04:02 PM Revision 4f0d170a (ceph): test: test_workload_gen: Change CLI option and add '--help' usage.
With this commit, we support the following options (and old ones are no
longer available):
--test-num-colls VAL ...
Joao Eduardo Luis
03:34 PM Revision 18d219e5 (ceph): rgw: replace dout with ldout
librgw can't use g_ceph_context
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
03:21 PM Feature #2222: osd: distinguish between 'degraded' and 'misplaced'
We should pick a designator that doesn't make it sound like the objects are lost. Greg Farnum
02:27 PM Feature #2222 (Resolved): osd: distinguish between 'degraded' and 'misplaced'
normal data migration happens with a acting set > the up set, so that we never drop below N replicas, but we still ca... Sage Weil
02:45 PM Feature #2087: lightweight filestore workload generator
Joao Eduardo Luis
02:07 PM Bug #2221 (Resolved): Monitor setup bugs
Carl reported several configuration issues when creating new monitors (based on the instructions at http://ceph.newdr... Greg Farnum
01:59 PM Revision a3bdf055 (ceph): test: test_workload_gen: Default arguments, and minor changes.
Besides adding support for default arguments, passed onto global_init(),
this commit fixes a conflict in Makefile.am,...
Joao Eduardo Luis
01:32 PM Revision 37cdbcd4 (ceph): log: fix up unittest
Fewer entries; compile.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
08:35 AM rgw Bug #2220 (Resolved): rgw: librgw dep on g_ceph_context
Fixed, commit:18d219e512a8e0f427a2229a71e15869cac3b593. Yehuda Sadeh
07:16 AM rgw Bug #2220 (Resolved): rgw: librgw dep on g_ceph_context
from last night's qa,... Sage Weil
04:37 AM Bug #2219: OSD's commit suicide with 0.44
I accidentally removed the core file(s) :(
Hope this one pops up again so I have a core file.
Wido den Hollander
04:11 AM Linux kernel client Tasks #2138: rbd: run xfstests on a local XFS filesystem over RBD
After setting up two rbd devices and making some fairly simple changes
to xfstests, then setting up appropriate envi...
Alex Elder
04:04 AM Linux kernel client Bug #2155: ceph: xattr: wrong value assumed for "no preferred PG"
This got rebased: 3489b42a72a41d477665ab37f196ae9257180abb
This has been sent as part of a pull request to Linus ...
Alex Elder
04:04 AM Linux kernel client Bug #2156: ceph: xattr: fix a possible buffer overrun bug
This got rebased: 3489b42a72a41d477665ab37f196ae9257180abb
This has been sent as part of a pull request to Linus ...
Alex Elder
04:03 AM Linux kernel client Bug #2157: ceph: xattr: fix nanosecond display on i_rctime
This got rebased: 3489b42a72a41d477665ab37f196ae9257180abb
This has been sent as part of a pull request to Linus ...
Alex Elder
04:01 AM Linux kernel client Bug #2064: ceph-client: messenger: nocrc flag not implemented correctly
It got rebased once more, and this should be the last:
37675b0f42a8f7699c3602350d1c3b2a1698a3d3
This has been s...
Alex Elder
03:52 AM Bug #2178: rbd: corruption of first block
Hi,
I decided to upgrade to "latest-n-greatest" in the test-cluster, to make sure, that if I hit the error again w...
Oliver Francke
02:58 AM Revision 94e3abf8 (ceph): Merge branch 'stable'
Sage Weil
12:22 AM Revision 8948ad01 (ceph): test: test_workload_gen: CodeStyle compliance and cleanup.
This commit aims at the compliance with Ceph's CodeStyle, as well
as cleaning up some lingering unused code.
Also, n...
Joao Eduardo Luis
12:22 AM Revision d172b40c (ceph): test: test_workload_gen: Destroy collections.
Joao Eduardo Luis
12:22 AM Revision 3770096a (ceph): test: test_workload_gen: Mimic an OSD's workload.
In it's current state, the workload generator will queue a lot of
transactions onto the FileStore, and will wait if n...
Joao Eduardo Luis
12:18 AM Revision 749826c2 (ceph): allow use of a separate journal block device
Samuel Just

03/27/2012

11:44 PM Revision ffc468f2 (ceph): osdmap: less noisy about osd additions during buildmap
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
11:44 PM Revision 36c2f27d (ceph): osdmaptool: fix clitest conf filename
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
11:37 PM Revision ca1f79b5 (ceph): dout: no newlines on dout_emergency
Preserve old behavior to avoid breaking all the cli tests.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
11:27 PM Revision d5360968 (ceph): throttle: fix off by one issue
We were blocking only if we exceeded max count, not if
we reached it.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdrea...
Yehuda Sadeh
11:23 PM Revision a52d048a (ceph): rgw: throttle incoming requests
Don't accept more than the number of threads, otherwise if cluster is
backed up for any reason we'd end up exhausting...
Yehuda Sadeh
11:16 PM Revision 30cadf01 (ceph): prebufferedstreambuf: fix typedef
'typename' not allowed here:
./common/PrebufferedStreambuf.h:27: error: using 'typename' outside of template
Signed...
Sage Weil
10:35 PM Revision 93ba4c00 (ceph): Merge branch 'wip-intent-fixes'
Yehuda Sadeh
10:35 PM Revision ca4fab47 (ceph): Merge branch 'master' of ssh://github.com/ceph/ceph
Yehuda Sadeh
10:35 PM Revision 16b60b3e (ceph): rgw: minor style fixes
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
10:12 PM Revision 4d74a7b2 (ceph): osd: fix handling of recovery sources when osds go down
If a source osd goes down, we need to
- reset any pulls (already did that before)
- remove peer from missing_loc s...
Sage Weil
10:03 PM Revision 8fdde24c (ceph): osd: remove down osds from peer_*_requested maps
This will leave less crap around to confuse recovery if a source osd goes
down and then up.
Signed-off-by: Sage Weil...
Sage Weil
10:02 PM Revision 1ee60873 (ceph): osd: maintain missing_loc_sources
This is a superset of all missing_loc values... everywhere we might
pull an object from, or are currently pulling fro...
Sage Weil
09:37 PM Revision 5dbb9715 (ceph): rgw: all intent log operations are now async
That includes removing a directory index object, and the removal of
the actual intent log object.
Signed-off-by: Yeh...
Yehuda Sadeh
09:20 PM Revision 0b1e3ed4 (ceph): osd: increase default heartbeat_interval to 6 seconds
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
09:12 PM Revision 69844496 (ceph): rgw: remove pool_list(), can't list_objects() on system buckets
pool_list() was broken, replaced now with pool_iterate(). list_objects()
shouldn't be used any more with system bucke...
Yehuda Sadeh
09:04 PM Revision 2e9079cf (ceph): rgw: intent log processing uses new pool_iterate()
intead of pool_list(), which is broken (assuming pgls results are
sorted, which are not).
Signed-off-by: Yehuda Sade...
Yehuda Sadeh
08:57 PM Revision 1814aac1 (ceph): Merge branch 'misc-fixes-for-review'
Greg Farnum
08:57 PM Revision d5c4015d (ceph): uclient: We want to release cache when we lose the CACHE cap, not gain it!
Looks like this was detected as a problem back in
84644dc56183b67050793a1b8da07850508b29d6 but the fix wasn't complet...
Greg Farnum
08:57 PM Revision c3b04644 (ceph): paxos: share_state sends every unknown value, including the stashed one
Sage points out that the stashed object might not be the same as the
one we actually archive. For instance, OSDMonito...
Greg Farnum
08:57 PM Revision 2acf4aea (ceph): mon: Paxos needs to store the latest version permanently on-disk.
Previously it was only storing this m->latest_value in the stash,
which of course got overwritten. And then when some...
Greg Farnum
08:57 PM Revision d0ba27ae (ceph): doc: add a short thing on kernel client troubleshooting.
I just noticed this sitting uncommitted in my tree.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Greg Farnum
08:02 PM Revision c89b7f22 (ceph): v0.44.1
Sage Weil
06:35 PM Revision 6044c5b8 (ceph): hadoop: define subsystem, fix logging
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
06:31 PM CephFS Bug #2218: CephFS "mismatch between child accounted_rstats and my rstats!"
The MDS log is at https://matthew.royhousehold.net/mds.a.log.1.gz (1505MB, md5 197ef232d50d27e2b7c2f62370c9c6b6) Matthew Roy
02:45 PM CephFS Bug #2218 (Need More Info): CephFS "mismatch between child accounted_rstats and my rstats!"
There's not enough info in the attached log to figure out what happened. I can tell you that your home directory beli... Greg Farnum
06:20 PM Revision ce61a83f (ceph): log: throttle message submission, trim recent
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
06:20 PM Revision fe56818e (ceph): config: configure log thresholds
- max new entries before we wait for flush
- max recent entries to keep around
Signed-off-by: Sage Weil <sage@newdre...
Sage Weil
06:05 PM Revision 339956df (ceph): log: don't spam -1 to syslog; add err_to_syslog for consistency
This matches the stderr settings.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:44 PM Revision 17a95c22 (ceph): log: use PrebufferedStreambuf
It's faster than ostringstream!
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:44 PM Revision bfa2bcd7 (ceph): prebufferedstreambuf: fix get_str()
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:43 PM Revision 0e3c0c44 (ceph): bench_log: flush
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:42 PM Revision 3a87e452 (ceph): log/EntryQueue: no implicit trim
dequeue() things explicitly if you want to remove them.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:41 PM Revision f66e0750 (ceph): utime_t: sprintf() method
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:41 PM Revision 6ab85264 (ceph): do_autogen: control optimization level
-O 2 -> -O2
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:41 PM Revision a4509273 (ceph): common: add PrebufferedStreambuf
Simple streambuf that uses a preallocated buffer, and then spills over
into a std::string if necessary.
Signed-off-b...
Sage Weil
05:41 PM Revision 23f0af3c (ceph): test log performance with PreallocatedStreambuf
- faster than ostringstream in optimistic case
- same as ostreamstream + std::string assignment in worst case (use
...
Sage Weil
05:41 PM Revision 8c5046fa (ceph): bench_log: simple util to time how long it takes to log stuff
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:41 PM Revision 362ca19b (ceph): log: move create_entry() into Log interface
This will let us be smarter than putting it on the heap.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:41 PM Revision c7242bfe (ceph): log: flush on_exit
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:41 PM Revision abfadb9b (ceph): assert: dump recent log entries on failed assertions
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:41 PM Revision f41887e3 (ceph): log: new logging infrastructure
- explicitly defined subsystems, and ceph_subsys_FOO enums to go with them
- modular log system with Entry object
- s...
Sage Weil
04:26 PM rgw Bug #2197 (Resolved): rgw: need to throttle incoming requests
Fixed, commit:a52d048ac429c3d2b6a9286d96253308f6588762. Yehuda Sadeh
04:10 PM Bug #2178: rbd: corruption of first block
The next step is to reproduce the corruption on the test cluster with logs:
debug osd = 20
debug ms = 1
debug...
Sage Weil
08:37 AM Bug #2178: rbd: corruption of first block
Well,
one more comment:
my guess would be, it has todo something with expansion of the "sparse-file" while writin...
Oliver Francke
05:24 AM Bug #2178: rbd: corruption of first block
Good morning ;)
meanwhile I have not been lazy. I've managed - with current setup in test-cluster - to produce "in...
Oliver Francke
04:07 PM Bug #2164: osd: scrub missing _, snapset attrs
wip-2164
it's a problem with the collection_move guard (or lack thereof)
Sage Weil
03:40 PM rgw Bug #2208 (Resolved): rgw: radosgw-admin temp remove failure
Fixed, merged at commit:93ba4c004a9269148a75b67da2522855cb1842a3. Yehuda Sadeh
02:19 PM Bug #2219 (Need More Info): OSD's commit suicide with 0.44
Can you look at the core file and 'thread apply all bt'? Sage Weil
05:57 AM Bug #2219: OSD's commit suicide with 0.44
... Wido den Hollander
05:03 AM Bug #2219 (Can't reproduce): OSD's commit suicide with 0.44
I noticed this myself today, but on IRC somebody else came along:... Wido den Hollander
02:03 PM Bug #2199 (Resolved): mon: get_bl osdmap_full/9583 No such file or directory
Merged to master in commit:1814aac17593dee0fa4c774d5b462f277f6698da, reviewed by Sage — even though I forgot to add t... Greg Farnum
12:25 PM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
Can you attach the full osd.1 log? Sage Weil
12:36 AM Bug #2211: osd: entity_inst_t OSDMap::get_inst(int) const
Over night I saw 16 OSD's go down with the same backtrace.
All OSD's were running with debug ms/osd set to 1, this...
Wido den Hollander
09:07 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
I've been off on other things, but this problem apparently recurred
even if the latest check-in (Josh's change) in p...
Alex Elder
08:38 AM CephFS Bug #2217: sync and O_DIRECT writes only write first extent in iov vector
The code should not be written that way.
However I think it doesn't matter at this point, because the only caller
...
Alex Elder

03/26/2012

11:48 PM Revision 974a2013 (ceph): objecter: don't call op_throttle_ops.take(1) unconditionally
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:17 PM Revision 679cd1fe (ceph): objecter: add in-flight ops throttling
In addition to ops length, we also want to throttle it by
actual number of ops.
Signed-off-by: Yehuda Sadeh <yehuda@...
Yehuda Sadeh
10:02 PM Revision d6b0cbd4 (ceph): config: use our assert
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
07:40 PM Revision c3dc6a6e (ceph): msg: assert pipe->msgr == msgr
Fixes: #2216
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
06:57 PM Revision e30b7710 (ceph): rbd: fix typo in default config
pyflakes would have caught this if 'all' weren't a built-in function Josh Durgin
06:43 PM Revision 483fcf80 (ceph): doc: include crush in toctree
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
06:30 PM Revision 3bd1f18e (ceph): doc: few notes on manipulating the crush map
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
06:24 PM CephFS Bug #2218 (Resolved): CephFS "mismatch between child accounted_rstats and my rstats!"
The mismatch is detected at 2012-03-26 18:39:54.306661... Matthew Roy
05:15 PM Revision 6db77158 (ceph): doc/dev/peering.rst: fix typo
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
04:39 PM Revision 1a0360cb (ceph): osd/: OpRequest is no longer a RefCountedObject, remove puts/gets
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
04:38 PM Revision ea377a08 (ceph): osd/: Convert OpRequest* to OpRequestRef
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
04:38 PM Revision 2cb6c7d0 (ceph): OSD: Add typedef for shared_ptr<OpRequest>
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
04:36 PM Revision 3ed784c9 (ceph): osd/: add mark_event to OpRequest and move tracking into OpTracker
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
03:51 PM Bug #2192: ceph-mon hangs consuming 100% CPU
It was reproduced all the time, for 0.44 also. After I adjusted cluster to have only one monitor problem has gone. (U... Vladimir Kulev
02:44 PM CephFS Bug #2217 (Resolved): sync and O_DIRECT writes only write first extent in iov vector
static ssize_t ceph_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t po...
Sage Weil
01:34 PM Bug #2199 (Fix Under Review): mon: get_bl osdmap_full/9583 No such file or directory
Re-pushed misc-fixes-for-review. Greg Farnum
09:59 AM Bug #2199 (In Progress): mon: get_bl osdmap_full/9583 No such file or directory
Sage pointed out the stash data structure isn't necessarily the same as the other stored data structures, so this nee... Greg Farnum
12:47 PM Messengers Cleanup #2216 (Resolved): SimpleMessenger should make sure it owns passed-in Connections
Sage Weil
10:50 AM Messengers Cleanup #2216 (Resolved): SimpleMessenger should make sure it owns passed-in Connections
Otherwise we get weird issues like #2212. Greg Farnum
12:38 PM Cleanup #2191: reexamine simple_spinlock
my log branch drops this for the dout logging. the last user is the buffer.h debugging (enabled manually via a macro... Sage Weil
12:06 PM RADOS Bug #2047: crush: with a rack->host->device hierarchy, several down devices are likely to cause b...
fwiw dropping the local search behavior fixes this bad behavior. the question is what probably was the local search ... Sage Weil
11:27 AM RADOS Bug #2047: crush: with a rack->host->device hierarchy, several down devices are likely to cause b...
Sage Weil
11:27 AM Bug #2210 (Duplicate): osd: some PGs remains remapped or degraded
this is actually a crush problem, see #2047. Sage Weil
09:45 AM Bug #2210: osd: some PGs remains remapped or degraded
#2173 has some osd logs and related info for the same problem on a less clean cluster. Thanks for the detailed steps ... Josh Durgin
10:36 AM CephFS Fix #2215 (Resolved): ceph-fuse does not invalidate page cache
Right now the userspace client doesn't invalidate the page cache when it loses the cache capability on an inode. Appa... Greg Farnum
09:58 AM Bug #2212 (Resolved): osd: FAILED assert(msgr->lock.is_locked())
ah, i was using wrong msgr, fixing! Sage Weil
05:50 AM Bug #2212 (Resolved): osd: FAILED assert(msgr->lock.is_locked())
With the new heartbeat code I noticed a couple of OSD's go down with:... Wido den Hollander
09:58 AM RADOS Bug #2214 (Resolved): crush: pgs only mapped to 2 devices with replication level 3
This is from #2173. Note that all 3 osds are up.... Josh Durgin
09:43 AM Bug #2173 (Resolved): MDS crash when start with end of buffer
Josh Durgin
06:04 AM Feature #2213 (Resolved): rbd: shouldn't need config file to get help
I just ran "rbd --help" on a pretty much un-configured machine and got:
global_init: unable to open config file.
...
Alex Elder
05:22 AM Bug #2211 (Resolved): osd: entity_inst_t OSDMap::get_inst(int) const
While trying out the new heartbeat code I encountered this crash:... Wido den Hollander
03:28 AM Revision e478a758 (ceph): vstart: enable omap for xattrs
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil

03/25/2012

08:39 PM Bug #2173: MDS crash when start with end of buffer
Shall we colse this bug, as the mds server was recovered by providing an empty session map and we can not reproduced ... soft crack
08:39 PM Bug #2210 (Duplicate): osd: some PGs remains remapped or degraded
Some PGs remains 'remapped' or 'degraded' status after adding an osd server.
The steps to re-produce the bugs:
1....
soft crack
03:05 PM Revision f4b2097a (ceph): Merge remote branch 'gh/wip-doc-peering'
Sage Weil
02:57 PM Revision d3bcac24 (ceph): Makefile: fix modules that cannot find pk11pub.h when compiling with NS...
Signed-off-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Jim Schutt
02:57 PM Revision 3ab28950 (ceph): don't override CFLAGS
leveldb adds -I flags to CFLAGS and CXXFLAGS, but if these macros are
overridden in the make command line, the flags ...
Alexandre Oliva
09:54 AM Feature #2087: lightweight filestore workload generator
Pushed a new commit to [1], making the code compliant with the CodeStyle and with Sage's suggestions on github.
[1...
Joao Eduardo Luis
04:47 AM Revision ef17c8c9 (ceph): add smoke suite
This could probably be collapsed into a bunch of singleton tasks to make
it simpler to track how many actual jobs res...
Sage Weil
04:20 AM Revision b5641ef3 (ceph): rgw: don't #include fcgi from rgw_common.h
ceph-dencoder #includes rgw_common.h, and needs to build even when
--without-radosgw is specified and libfcgi isn't i...
Sage Weil
04:09 AM Revision 1c1192a9 (ceph): backfill: use 'rbd' pool instead of 'data'
(data has a replay interval, which makes writes take longer to resume
after repeering)
Sage Weil
04:09 AM Revision 397e7f2f (ceph): add osd_recovery task to test divergent osd logs
Sage Weil

03/24/2012

11:07 PM Revision 24910c3b (ceph): add osd-recovery test
Sage Weil
11:07 PM Revision 6bf9c957 (ceph): renamed backfill -> osd_backfill
Sage Weil
11:05 PM Revision ca9a5a4a (ceph): rename backfill -> osd_backfill
Sage Weil
10:36 PM Revision 22e80874 (ceph): put filestore xattr option in [global]
...for test_filestore_idempotent's benefit Sage Weil
09:41 PM Feature #2134: qa: smoke suite
Sage Weil
09:04 PM Feature #1802 (Resolved): qa: test to exercise divergent osd logs
Sage Weil
03:10 PM Bug #2192: ceph-mon hangs consuming 100% CPU
Is this reproducible? Are you able to connect to the ceph-mon process with gdb? Sage Weil
03:06 PM Bug #2185 (Won't Fix): osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_ran...
Sage Weil
08:13 AM Feature #2087: lightweight filestore workload generator
Pushed a working version to ceph's git repository, branch wip-2087 [1]. Feedback would be appreciated.
[1] - https...
Joao Eduardo Luis

03/23/2012

08:27 PM Revision 2ec8f27f (ceph): rados_bench: generate_object_name now takes a buffer length
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Samuel Just
05:27 PM Bug #2209 (Resolved): osd: read kb stats not tracked?
Sage Weil
01:21 PM Bug #2196: `rados bench` will write test objects with a constant oid, under-reporting performance.
2ec8f27f58adca40d125051a23547b639ee7d5f6 Samuel Just
01:21 PM Bug #2196 (Resolved): `rados bench` will write test objects with a constant oid, under-reporting ...
Samuel Just
12:53 PM rgw Bug #2208 (Resolved): rgw: radosgw-admin temp remove failure
The radosgw-admin temp remove on congress goes into infinite loop when trying to list the .intent-log pool. Yehuda Sadeh
11:07 AM Bug #2200 (Can't reproduce): mon: not accepting new connections
Yehuda's indicated that this might be tied in to networking issues that were ongoing at the time. Given the symptoms ... Greg Farnum
11:04 AM Bug #2199 (Fix Under Review): mon: get_bl osdmap_full/9583 No such file or directory
I believe this is fixed in misc-fixes-for-review commit:e08b489d094efe384c3db639af0be765665bee23. Sage needs to revie... Greg Farnum

03/22/2012

11:09 PM Bug #2200: mon: not accepting new connections
Okay, that appears to not be it (the connections established and terminated match for clients and are only off by 9 o... Greg Farnum
10:09 PM Bug #2200: mon: not accepting new connections
There's not a lot I can do to diagnose this with just logs; the Monitors don't refuse connections like that on their ... Greg Farnum
09:42 AM Bug #2200 (Can't reproduce): mon: not accepting new connections
Following a networking downtime and monitors restart (as described in #2199), and following a recovery process, all a... Yehuda Sadeh
10:00 PM Bug #2199 (In Progress): mon: get_bl osdmap_full/9583 No such file or directory
Looks like the problem is that the Monitor got elected leader, and while it collected all the state it didn't write i... Greg Farnum
10:00 AM Bug #2199: mon: get_bl osdmap_full/9583 No such file or directory
My guess/hope is that this is one of the issues solved by the monitor slurp and other fixes since 0.41, but I haven't... Greg Farnum
09:41 PM Revision 21a170e8 (ceph): doc: dev/peering.rst edits from Greg
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:12 PM Bug #2207 (Resolved): osd: crash when op length is greater than op input data
This could happen due to a malicious or buggy client. I caused this with an accidentally empty request, with positive... Josh Durgin
05:10 PM CephFS Documentation #2206 (Resolved): Need a control command to gracefully shutdown an active MDS prior...
There is currently no way to gracefully shutdown an active MDS and allow a standby to activate or to transfer the act... Matthew Roy
04:53 PM Bug #2205 (Won't Fix): mkcephfs throws "No such file or directory" errors when the pwd the script...
When executing mkcephfs on a new cluster the script throws the message "bash: line 0: cd: /home/matthew/forCeph: No s... Matthew Roy
03:33 PM Revision 8fa904a6 (ceph): doc: update dev/peering document
- fix discussion of last epoch started
- define terms for current and past intervals
- describe role of pg info
- rem...
Sage Weil
02:55 PM Revision de867632 (ceph): msgr: fix tcp.cc linkage
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
02:53 PM Revision fd9935b7 (ceph): cephtool: don't prefix log items
This just makes it hard to read them.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
02:46 PM Subtask #2201: Document old design
+1; I have no idea what this bug is for Dan Mick
11:39 AM Subtask #2201: Document old design
Old design of...what?
(I see now that it's connected to the omap stuff, but if you could include a little more con...
Greg Farnum
11:34 AM Subtask #2201 (In Progress): Document old design
Samuel Just
11:33 AM Subtask #2201 (In Progress): Document old design
Samuel Just
01:08 PM Bug #2196 (In Progress): `rados bench` will write test objects with a constant oid, under-reporti...
Samuel Just
11:34 AM Subtask #2204 (Rejected): implement upgrade from old design to new design
Samuel Just
11:33 AM Subtask #2203 (In Progress): implement new design
Samuel Just
11:33 AM Subtask #2202 (Rejected): Document new design
Samuel Just
11:33 AM Feature #2149 (In Progress): osd: use omap for snap collections
Samuel Just
11:17 AM Feature #2198: add an option to force a down osd to be marked immediately out
Hmm, yeah, I forgot about that.
Somebody was asking about it; I'm not sure if they cared exactly but I'm sure there ...
Greg Farnum
11:08 AM Feature #2198: add an option to force a down osd to be marked immediately out
Not really, a write will still go to N-1 replicas until the new one is backfilled up through the object's position.
...
Sage Weil
11:00 AM Feature #2198: add an option to force a down osd to be marked immediately out
It guarantees that you always have the set number of copies on-disk when you get a commit, instead of probably having... Greg Farnum
10:47 AM Feature #2198: add an option to force a down osd to be marked immediately out
What's the motivation for doing that? Is it any better than setting the out interval to be something very short? Sage Weil
09:14 AM Bug #2116: Repeated messages of "heartbeat_check: no heartbeat from"
see new wip-osd-hb branch
Sage Weil

03/21/2012

11:41 PM Revision 2e21adf2 (ceph): Objecter: resend linger_ops on any change
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Samuel Just
11:35 PM Revision b47454b6 (ceph): ObjectStore: add COLLECTION_MOVE to dump
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
11:35 PM Revision 23313ee6 (ceph): FileStore: whitelist COLLECTION_MOVE on replay
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
11:35 PM Revision ec52eeb2 (ceph): FileStore: remove src on EEXIST during collection_move replay
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
11:35 PM Revision 52aff487 (ceph): ObjectStore: Add collection_move to generate_instances
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
09:01 PM Revision 3caa4319 (ceph): ceph: define and use a shell_scripts Makefile variable
Define a variable "shell_scripts" in the Makefile.in, and use it
along with some pattern rules to avoid some duplicat...
Alex Elder
09:01 PM Revision 1b2a0669 (ceph): ceph-kdump-copy: add tools for saving kdumps
This puts in place an init script and a command it runs to save a
kernel core dump to a remote server when a panic or...
Alex Elder
08:41 PM Bug #2199: mon: get_bl osdmap_full/9583 No such file or directory
kept logs for the failing monitor under /var/log/ceph/2199 Yehuda Sadeh
08:26 PM Bug #2199 (Resolved): mon: get_bl osdmap_full/9583 No such file or directory
Happened on congress (afair, off 0.41). One monitor is out for more than a month. Following network outage, both moni... Yehuda Sadeh
07:00 PM Revision 6f0f250b (ceph): suite: add missing print statement
Josh Durgin
06:58 PM Revision 8a9a5670 (ceph): suite: fix print statement when summary doesn't exist
Josh Durgin
04:59 PM Feature #2198 (New): add an option to force a down osd to be marked immediately out
Samuel Just
02:25 PM rgw Bug #2197 (Resolved): rgw: need to throttle incoming requests
In case we can't handle requests, we'd end up accepting requests indefinitely thus we consume fds endlessly. This wil... Yehuda Sadeh
01:30 PM Revision d0e8f148 (ceph): doc: update list of debian dists
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:28 PM Revision a608a8fe (ceph): Merge branch 'stable'
Sage Weil
12:52 PM Bug #2196 (Resolved): `rados bench` will write test objects with a constant oid, under-reporting ...
(As discussed on @#ceph@, 2012/03/21 -- with thanks to @joshd@)
The command @rados bench@ generates a sequence of ...
David McBride
08:21 AM Bug #2178: rbd: corruption of first block
The next object is whatever the MBR points to. You can find the object name from the sector offset that gdisk gives y... Josh Durgin
02:55 AM Bug #2178: rbd: corruption of first block
Hi Josh,
thanks for taking the time to investigate this... And yes, many others show the same behaviour. Is "the n...
Oliver Francke
06:50 AM Feature #2127: Save kernel core dumps on all of our test machines
I seem to remember seeing a reference to 'mkcrashrd', a mkinitrd type script that generates the initrd image the cras... Sage Weil
02:00 AM Revision 91c08f6e (ceph): Add watch op to rados.py
Signed-off-by: Samuel Just <sam.just@dreamhost.com> Samuel Just
12:51 AM Revision 72361784 (ceph): Objecter: resend linger_ops on any change
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Samuel Just
12:51 AM Revision 3019d460 (ceph): TestRados: Add watch
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>
Samuel Just
12:20 AM Revision 2998368a (ceph): rgw: remove unused definition
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:20 AM Revision 4760536f (ceph): rgw: keep pool placement info also in cacheable location
Mirror the pools placement info, so that we can cache it.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
12:00 AM Revision f1563a66 (ceph): Revert "Objecter: add op->resend_on_any_change"
This reverts commit c53194d75390dd6d5aa4a9a33f741cbd106e3338.
recalc_linger_op_target is used for linger_ops
Signed...
Samuel Just

03/20/2012

11:11 PM Revision 2daff0e9 (ceph): ReplicatedPG: osd_max_notify_timeout -> osd_default_notify_timeout
This setting should not override user specified timeout.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Samuel Just
11:11 PM Revision c53194d7 (ceph): Objecter: add op->resend_on_any_change
lingers must be resent even if the primary does not change.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Samuel Just
11:11 PM Revision fc7a1bda (ceph): ReplicatedPG: return -EBUSY on delete for objects with watchers
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
10:15 PM Revision 6a5cbec3 (ceph): rgw: replace bucket_id generation
bucket_id is now string: <global instance id>.<num> where
num is increasing monotonically within the current rgw
inst...
Yehuda Sadeh
09:07 PM Feature #2127 (In Progress): Save kernel core dumps on all of our test machines
I finally have crash dumps getting packaged and sent over to a
remote machine reliably. The problem is that it does...
Alex Elder
06:59 PM Bug #2178: rbd: corruption of first block
I looked at the block you attached, and compared it to the first 4MiB of my desktop's hard drive. It looks like it co... Josh Durgin
03:58 AM Bug #2178: rbd: corruption of first block
Hi *,
any update on this topic? Cause we are working for hours and days with three people to rescue as many images...
Oliver Francke
06:41 PM Revision cdd5298d (ceph): v0.44
Sage Weil
05:59 PM Revision e42fbb70 (ceph): rgw: process default alt args before processing conf file
this fixes #2189
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
05:52 PM Revision e0b8f7a0 (ceph): rgw: process default alt args before processing conf file
this fixes #2189
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
05:37 PM Revision 51a07339 (ceph): rgw: incrase socket backlog
20 is too small
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
05:25 PM Revision 5b331987 (ceph): rgw: fix internal cache api
This fixes issue #2190
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
05:14 PM rgw Bug #2193 (Resolved): rgw: .pools.avail is not cached
Fixed, commit:4760536fe573c702bac8fb1d51213d76059e32dc.
We now mirror the info in the object. Still keeping the om...
Yehuda Sadeh
09:28 AM rgw Bug #2193 (Resolved): rgw: .pools.avail is not cached
Probably due to recent omap changes, we don't cache omap operations. Either we cache it, or just keep available pools... Yehuda Sadeh
03:10 PM rgw Feature #2194 (Resolved): rgw: replace bucket-marker-ver with better, fast, more scalable solution
Fixed, commit:6a5cbec38b761d524e699e2a7410a340d093ccca. Yehuda Sadeh
09:32 AM rgw Feature #2194 (Resolved): rgw: replace bucket-marker-ver with better, fast, more scalable solution
We use this object in order to create unique prefix for bucket objects (we do it at bucket creation). Instead of this... Yehuda Sadeh
02:50 PM Revision 815fc3e2 (ceph): suite: failed runs might not have durations
This was one cause of emails not being sent - stale /tmp/cephtest dirs
fail without recording a duration.
Josh Durgin
10:47 AM rgw Bug #2189 (Resolved): rgw: can't change debug level through ceph.conf
Fixed, commit:e0b8f7a0331b0ceee54a911bb9231cb168eb2d0f. Yehuda Sadeh
10:28 AM rgw Bug #2190 (Resolved): rgw: cache disabled
Fixed, commit:5b3319870ea9d6c715c671e006e3a772008e3e78. Yehuda Sadeh
09:43 AM CephFS Feature #2195 (Resolved): Allow removal of last MDS if there's no filesystem
Right now you can't remove the last MDS from your cluster, which means that if you aren't using it and it's off you w... Greg Farnum
05:43 AM Bug #2192 (Won't Fix): ceph-mon hangs consuming 100% CPU
I have a test setup of two nodes each running 0.43 mds, mon and osd. I mount ceph kernel filesystem at /srv/ceph on b... Vladimir Kulev

03/19/2012

11:36 PM Cleanup #2191 (Resolved): reexamine simple_spinlock
We've got a homebrewed spinlock implementation in src/common/simple_spin.h/cc. It was written so we could use dout in... Greg Farnum
11:10 PM Revision f923b840 (ceph): OSD: do not hold obc lock in disconnect_session_watches
ObjectContext::lock is used only for implementing read_lock and
write_lock. PG::lock is used to protect the ObjectCo...
Samuel Just
09:16 PM Revision a65d4136 (ceph): suite, coverage: use absolute dirs for isdir checks
This fixes the results to wait for all jobs to complete again. Josh Durgin
06:57 PM Revision bdb72c28 (ceph): filestore_idempotent: get coverage and coredumps
Josh Durgin
06:31 PM Revision 6c8db1a8 (ceph): suite: more results logging
Josh Durgin
05:34 PM rgw Bug #2190 (Resolved): rgw: cache disabled
in master branch only, due to internal api change. Yehuda Sadeh
05:33 PM rgw Bug #2189 (Resolved): rgw: can't change debug level through ceph.conf
Yehuda Sadeh
05:12 PM Bug #2188 (Resolved): mon: mds rm should be harder to break things with
If you run ceph mds rm 0 on a healthy cluster, it breaks the Monitor's world. I'm uncomfortable with the command exis... Greg Farnum
04:04 PM Bug #2183 (Resolved): osd: lockdep cycle with obc lock and watch_lock
pushed to master f923b840edec79df5791a7fb7fdec8b0b40f25f1 Samuel Just
03:33 PM Bug #2183: osd: lockdep cycle with obc lock and watch_lock
I believe it's inappropriate to hold obc->lock there anyway, pg lock serves that purpose. Samuel Just
11:07 AM Bug #2185: osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_range()
In the wip-rbd-bid branch that I pushed last week I added an option to the rbd tool to create images using existing d... Yehuda Sadeh
11:01 AM Bug #2185: osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_range()
should be pretty easy to rebuild the xattr, removing the object would corrupt the rbd image Samuel Just

03/18/2012

10:36 PM Bug #2173: MDS crash when start with end of buffer
I have managed to start mds server after resetting the journal. So I can get my data back.
Thanks very much to all o...
soft crack
06:56 PM Revision 7173a8af (ceph): ceph.conf: no comment
Sage Weil
06:06 PM Revision 7de798f6 (ceph): ceph.conf: set 'filestore xattr use omap = true'
Sage Weil
05:50 PM Revision 7d2e1056 (ceph): fix teuthology-ls isdir check
Sage Weil
05:48 PM Revision 94f0ba1e (ceph): run valgrind with cwd set to /tmp/cephtest/archive/coredump
This lets us capture the vgcore.* files, which always go to valgrind's
cwd.
Fixes: #1953
Sage Weil
04:09 PM Revision fd851304 (ceph): ReplicatedPG: there should be no object_contexts during on_activate
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
04:08 PM Revision 6c17a7b3 (ceph): Merge branch 'next'
Sage Weil
04:08 PM Revision 77c08f86 (ceph): osd: fix object_info.size mismatch file due to truncate_seq on new object
If the first write that creates an object includes a truncate_seq and
truncate_size, we were taking the truncte patch...
Sage Weil
01:46 PM CephFS Bug #2187 (Can't reproduce): pjd chown/00.t failed test 97
on both ceph-fuse and kclient, nightly_coverage_2012-03-17-a,
> 1727 FAIL scheduled_teuthology@teuthology collection...
Sage Weil
01:43 PM CephFS Bug #2159 (Resolved): ceph-fuse: big_writes option not recognized
Sage Weil
12:09 PM Bug #2080 (Resolved): osd: scrub on disk size does not match object info size
Sage Weil
12:09 PM Bug #1953: teuthology: core files aren't archived when using valgrind
Sage Weil
12:07 PM Bug #2164: osd: scrub missing _, snapset attrs
this was non-btrfs, right after the new idempotent replay stuff was fixed. Sage Weil
10:50 AM Bug #2186 (Can't reproduce): osd: shutdown race
... Sage Weil
10:07 AM Bug #2180 (Resolved): osd/ReplicatedPG.cc: 3381: FAILED assert(obc->watchers.size() == 0)
Sage Weil

03/16/2012

11:59 PM Revision 619fe730 (ceph): .gitignore: xattr_bench
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
10:06 PM Revision 3a6c085e (ceph): heartbeatmap: use utimes(2) instead of futimens(2)
For poor users with ancient glibc. We don't much care about rename races
here anyway.
Signed-off-by: Sage Weil <sag...
Sage Weil
09:36 PM Revision 63ec06b3 (ceph): osd: remove special handline for head recovery from clone
This breaks because:
- we don't have the head or current snapset
- get_object_context() creates a new snapset, whi...
Sage Weil
08:49 PM Revision d8bcc1b3 (ceph): config: fix recursive locking of md_config_t::lock
Signed-off-by: Sage Weil <sage@newdream.net>
Reviewed-by: Greg Farnum <gregory.farnum@dreamhost.com>
Sage Weil
08:30 PM Revision 58c5d5a0 (ceph): osd: ReplicatedPG::create_object_context()
New helper that creates a new object context.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
08:30 PM Revision d4addf57 (ceph): osd: re-use create_object_context() in get_object_context()
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
08:30 PM Revision 15d85af4 (ceph): osd: explicitly create new object,snap contexts on push
We specifically want to use this during recovery to avoid loading the obc
or ssc for a previous version of the object...
Sage Weil
08:28 PM Revision 01924a22 (ceph): disable rbd thrash workload, #2174
Sage Weil
08:04 PM Revision 96780bd1 (ceph): osd: create_snapset_context()
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
07:15 PM Revision 872bdd0d (ceph): osd: ensure we don't clobber other *contexts when registering new ones
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
06:52 PM Revision 9791035d (ceph): Merge branch 'wip_omap_xattrs'
Samuel Just
06:44 PM Revision 07b97fe7 (ceph): suite: log results and coverage generation
Need to figure out where and when results emails are failing. Josh Durgin
06:40 PM Revision 2a593dda (ceph): RadosModel: test xattrs with omap
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:40 PM Revision a49a1972 (ceph): ReplicatedPG,FileStore: clone should copy xattrs as well
_make_clone (called from make_writeable) and _rollback_to included
attr reads from head or a clone. In that case, an...
Samuel Just
06:40 PM Revision 14506dc6 (ceph): FileStore: add support for omap xattrs
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:31 PM Revision a5f143d2 (ceph): Merge branch 'wip-msgr4'
Reviewed-by: Greg Farnum <gregory.farnum@dreamhost.com> Sage Weil
06:29 PM Revision 983fd190 (ceph): ObjectMap: add interface for storing xattrs
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:29 PM Revision d8325e50 (ceph): DBObjectMap: implement xattr interface
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:29 PM Revision fdb92748 (ceph): test_object_map: update unit test for xattr
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:29 PM Revision 8fc43179 (ceph): config_opts.h: opts for omap_xattrs
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:29 PM Revision ecd875fe (ceph): tests/: Added xattr bench
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:29 PM Revision b09fb15d (ceph): ObjectMap: use Index object for locking rather than path object
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:28 PM Revision 9fd4a12a (ceph): DBObjectMap: add support for storing xattrs
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
06:18 PM Bug #2185: osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_range()
strace indicated we had a missing xattr on
2268 stat("/data/osd0/current/164.2_head/rb.0.0.000000000000__head_DA6...
Sage Weil
06:02 PM Bug #2185: osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_range()
... Sage Weil
03:33 PM Bug #2185: osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_range()
Here output from osd.3 after recent crash:
root@fcmsnode3:/data/osd3/current# find 0.0_head
0.0_head
0.0_head/10...
Oliver Francke
03:22 PM Bug #2185 (Won't Fix): osd/ReplicatedPG.cc: 5938: FAILED assert(r >= 0) in ReplicatedPG::scan_ran...
... Sage Weil
06:01 PM Bug #2173: MDS crash when start with end of buffer
Talked more on irc, soft crack is trying to reset his journal since it looks like at least all his metadata objects a... Josh Durgin
04:43 PM Bug #2173: MDS crash when start with end of buffer
osd map file for 'ceph osd getmap 3212 -o /tmp/osdmap' soft crack
01:31 PM Bug #2173: MDS crash when start with end of buffer
Greg: look at the osd dump above: all pools are rep size 3. Josh Durgin
01:13 PM Bug #2173: MDS crash when start with end of buffer
Did all the pools get set to 3x replication, or are the confused PGs all part of the metadata pool? Greg Farnum
12:26 PM Bug #2173: MDS crash when start with end of buffer
Could you attach the output of 'ceph osd dump 3212' and the binary version of that osdmap (ceph osd getmap 3212 -o /t... Josh Durgin
09:56 AM Bug #2173: MDS crash when start with end of buffer
Unfortunately we can see that this assert too is caused by ENOENT on an object that really ought to be there, which m... Greg Farnum
08:19 AM Bug #2173: MDS crash when start with end of buffer
Can you post an mds log with debug mds = 20 leading up to that last crash?
Resetting the journal is not something ...
Sage Weil
08:15 AM Bug #2173: MDS crash when start with end of buffer
I managed to insert a empty sessionmap. The server continue starting.
And I get an assert error:...
soft crack
05:39 PM Revision 0904c7b7 (ceph): configure: fix warnings
Finally!
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
05:04 PM Revision f2e6b8d7 (ceph): ReplicatedPG: populate_object_context during handle_pull_response
A cached objectcontext should always have its watchers populated.
Signed-off-by: Samuel Just <samuel.just@dreamhost....
Samuel Just
04:43 PM Revision 4cfc34f8 (ceph): leveldb: .gitignore TAGS
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
04:40 PM Revision 5db6902b (ceph): leveldb: un-revert
Accidentally reverted by c2af646b38995ba005140e748a21baba4263e53f.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
02:33 PM Bug #2080: osd: scrub on disk size does not match object info size
wip-2080 Sage Weil
01:33 PM Bug #2184 (Resolved): audit calls to populate_obc_watchers and add watch/notify to RadosModel
Samuel Just
01:32 PM Feature #2125 (Resolved): osd: put large xattrs in leveldb
Samuel Just
01:20 PM Bug #2183: osd: lockdep cycle with obc lock and watch_lock
crashed it with this mutl... Sage Weil
01:18 PM Bug #2183 (Resolved): osd: lockdep cycle with obc lock and watch_lock
... Sage Weil
12:04 PM Bug #2180: osd/ReplicatedPG.cc: 3381: FAILED assert(obc->watchers.size() == 0)
Hi Sage,
here the according log after upgrading and starting 0.43-1...
Hope it helps,
Oliver.
Oliver Francke
08:56 AM Bug #2180 (Resolved): osd/ReplicatedPG.cc: 3381: FAILED assert(obc->watchers.size() == 0)
... Sage Weil
10:58 AM Bug #2182 (Resolved): audit osd reads for reads from potentially unstable objects
In particular, there are places we read object_info and snapset outside of the get_object_context and get_snapset_con... Samuel Just
10:50 AM Bug #2181 (Won't Fix): 4051: FAILED assert(!missing.is_missing(soid)) in ceph version 0.43-244-g9...
v0.43 and this commit from master aren't compatible; the final v0.44 will have a protocol rev to prevent this problem. Sage Weil
10:21 AM Bug #2181 (Won't Fix): 4051: FAILED assert(!missing.is_missing(soid)) in ceph version 0.43-244-g9...
Hi Sage,
here u r. This was the version, which failed, too, after all others didn't help either... Similar with al...
Oliver Francke
09:13 AM Bug #2132: FAILED assert(!missing.is_missing(soid))
Oliver Francke wrote:
> Well,
>
> its tagged as resolved, but today another node died...:
>
> osd/ReplicatedPG...
Sage Weil
05:36 AM Bug #2132: FAILED assert(!missing.is_missing(soid))
Well,
its tagged as resolved, but today another node died...:
osd/ReplicatedPG.cc: In function 'void Replicated...
Oliver Francke
04:19 AM Bug #2178: rbd: corruption of first block
Here is one of many, where the header is missing:
--- 8-< ---
fcms@fcmsnode3:~$ rbd ls 1320396354
vm-451-disk-1....
Oliver Francke
12:34 AM Revision 8fbd087d (ceph): results: make sure email is sent before anything else fails
Josh Durgin

03/15/2012

06:08 PM Bug #2173: MDS crash when start with end of buffer
Sorry for mistake.
ceph osd dump -o -:
2012-03-16 09:10:04.887611 mon <- [osd,dump]
2012-03-16 09:10:04.888161...
soft crack
06:01 PM Bug #2173: MDS crash when start with end of buffer
ceph -s:... soft crack
10:43 AM Bug #2173: MDS crash when start with end of buffer
Well that's exciting; this means it's an OSD bug.
The meaning of that output is that of your 209 PGs, 185 are happy;...
Greg Farnum
05:35 PM Revision 89ccd95a (ceph): osd: maybe clear DEGRADED on recovery completion
We set degraded if we don't have enough "active" replicas, which excludes
the backfill target. We need to recheck th...
Sage Weil
05:32 PM Revision b4572351 (ceph): Revert "disable rbd thrash workload, #2174"
This reverts commit 1bec416c7c7ff8a6462d94baaba8e7da73e88ab4.
Fixed with #2174
Sage Weil
12:58 PM rgw Feature #1941 (Rejected): rgw: revisit bucket removal
Sage Weil
12:57 PM rgw Feature #785 (Rejected): rgw: fix filesystem backend
Sage Weil
10:29 AM Bug #2160 (Resolved): active+recovering+degraded+backfill becomes active+clean+degraded when reco...
Sage Weil
09:49 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
The test that reproduced the problem has now run once to completion
without hitting it. Therefore it's ready to shi...
Alex Elder
08:35 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
... Alex Elder
07:57 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
Thanks Alex. I remember thinking it fixed a race initially, but then going back later and being unable to find the ra... Josh Durgin
07:43 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
That's excellent Josh. I'll use it, it's basically what I was
thinking of doing anyway, now I'll just use yours. D...
Alex Elder
07:38 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
That analysis of the race looks correct to me. The first unapplied patch in wip-rbd would have fixed this (9a3e22a0ce... Josh Durgin
07:14 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
... Alex Elder
07:12 AM Linux kernel client Bug #2174: rbd: iozone thrashing failure
I think I can explain this:
[ 265.117432] INFO: trying to register non-static key.
[ 265.149933] the code is ...
Alex Elder
12:16 AM Revision 826d30f1 (ceph): rgw: remove extra layer of RGWAccess
Not needed, now that we got rid of RGWFS
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh

03/14/2012

11:33 PM Revision 80e2a5e8 (ceph): msgr: switch all users over to abstract interface
This will let us transparently swap implementations out.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
11:29 PM Revision 1e1453c1 (ceph): msgr: introduce static Messenger::create() function
Create a new messenger, with whatever implementation is appropriate.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
11:29 PM Revision d26feffd (ceph): msgr: promote more methods to abstract Messenger interface
This will be everything that people actually use.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
11:01 PM Revision c2af646b (ceph): rgw: put_obj() uses bufferlist instead of extra alloc/copy
makes it cleaner.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
11:01 PM Revision 2b3bfd0c (ceph): rgw: remove fs backend
was broken anyway
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
10:51 PM Revision 1bec416c (ceph): disable rbd thrash workload, #2174
Sage Weil
08:53 PM Linux kernel client Bug #2174: rbd: iozone thrashing failure
I tried reproducing the problem, and although I'm not sure I know
how to recognize it my test did end in failure.
...
Alex Elder
09:54 AM Linux kernel client Bug #2174 (Can't reproduce): rbd: iozone thrashing failure
consistently failing
- ceph:
log-whitelist:
- wrongly marked me down or wrong addr
- objects unfo...
Sage Weil
08:32 PM Revision e14d428c (ceph): Merge branch 'master' of github.com:ceph/teuthology
Mark Nelson
08:32 PM Revision 2b879905 (ceph): Merge branch 'master' of github.com:ceph/teuthology
Mark Nelson
08:01 PM Revision a81b23e2 (ceph): Merge branch 'next'
Sage Weil
07:59 PM Revision bec47b57 (ceph): introduce CEPH_FEATURE_OMAP
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
07:55 PM Revision 8c96fd26 (ceph): leveldb: new .gitignore entry
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
07:14 PM Revision 20d11714 (ceph): osd: rev cluster internal protocol
This covers:
- the push/pull changes in 0.43 (which we forgot to protect against; see
#2132)
- the new omap stuff ...
Sage Weil
06:23 PM Bug #2173: MDS crash when start with end of buffer
Thanks for your responses.
I created this ceph file system with 1 mon, 1 osd, 1 mds. It works perferctly, and I wr...
soft crack
04:11 PM Bug #2173: MDS crash when start with end of buffer
Huh. Is this a new filesystem? Have you had any problems with the RADOS cluster (the OSDs)?
What's happening now i...
Greg Farnum
04:16 AM Bug #2173: MDS crash when start with end of buffer
I also tried: 'ceph-mds -i 1 -d --reset-journal 0'.
It just freeze.
soft crack
04:14 AM Bug #2173 (Resolved): MDS crash when start with end of buffer
My system is ubuntu 11.10 64bit. Mds just crashes when startup.
I noticed the message: 'No such file or directory'...
soft crack
04:36 PM Revision a0bcab5a (ceph): ceph-fuse: make big_writes optional via 'fuse big writes'
Fixes: #2159
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
04:35 PM rgw Bug #2001 (Resolved): radosgw memory leak
At this point I can't see any other leak (I already fixed one). Doesn't mean that there isn't another one, but I'm re... Yehuda Sadeh
04:08 PM CephFS Bug #2179 (Resolved): mds: don't crash on nonexistent SessionMap
Inspired by #2173. When the MDS tries to load the SessionMap it unconditionally decodes it, which causes a crash if t... Greg Farnum
03:46 PM Feature #2127: Save kernel core dumps on all of our test machines
http://linux.die.net/man/8/netdump
this mechanism looks simpler?
Sage Weil
02:55 PM Feature #2127: Save kernel core dumps on all of our test machines
Wed Mar 14 11:14:50 CDT 2012
OK, I got kernel core dumps and crash working in Ubuntu 11.10.
A lot of what I use...
Alex Elder
02:54 PM Feature #2127: Save kernel core dumps on all of our test machines
Oh, I forgot to mention I also wrote a little program that extracts
identifying information from a dump file that "k...
Alex Elder
02:48 PM Feature #2127: Save kernel core dumps on all of our test machines
I have been able to generate a core dump on an Ubuntu system.
I have transferred the result using scp to another hos...
Alex Elder
02:16 PM Bug #2178 (Resolved): rbd: corruption of first block
Sage Weil
01:02 PM Bug #2132 (Resolved): FAILED assert(!missing.is_missing(soid))
Sage Weil
12:01 PM Bug #2132: FAILED assert(!missing.is_missing(soid))
Aha, that explains it... the 0.42.2 and 0.43 interaction looks like the culprit here. We should have made them expli... Sage Weil
11:37 AM Bug #2132: FAILED assert(!missing.is_missing(soid))
All cephfs workload. It could be a versioning issue, I don't have the syslogs anymore that would show when I updated ... Matthew Roy
11:19 AM Bug #2132 (Need More Info): FAILED assert(!missing.is_missing(soid))
Matthew Roy: What was the nature of the workload? rbd? ceph fs? Samuel Just
11:06 AM Bug #2132: FAILED assert(!missing.is_missing(soid))
Josh Durgin wrote:
> stxShadow saw this as well.
It looks like in stxshadow's case, it was a version mismatch (cr...
Sage Weil
11:23 AM CephFS Cleanup #2177 (Resolved): mds: play nicely with omap
Convert the MDS to use OMAP properly.
There is at least one specific thing: right now it has optimizations for whe...
Greg Farnum
10:46 AM Bug #2176 (Resolved): dependencies not checked by autoconf
I recently resurrected a build of the user-mode and kernel clients on CentOS and found that I was missing a few packa... Anonymous
10:09 AM rgw Feature #2171: rgw: asynchronously calculate md5
Actually, I think it'll be easier doing it the other way around. As we already write the object asynchronously we can... Yehuda Sadeh
10:08 AM Documentation #2175 (Resolved): doc: fix doc build errors
e.g., http://ceph.newdream.net/gitbuilder-doc/log.cgi?log=a0bcab5a583e6c1fd87430252590ec902d1b6b98
It would be gre...
Sage Weil
09:56 AM Bug #2022: osd: misdirectect request
Just saw this with a different workload:... Sage Weil
09:51 AM CephFS Bug #2071: kclient: pjd mkfifo failures
hit this again:... Sage Weil
09:49 AM rgw Cleanup #2166 (Resolved): rgw: make sure librgw doesn't link against libfcgi
Fixed, commit:e19417ef55c713e60c61edd0de7c2228953407a1. Yehuda Sadeh
09:48 AM rgw Bug #2170 (Resolved): librgw references g_ceph_context
Fixed, commit:5912312c14a6214f4318fd7bfb6fd08714458b6f. Yehuda Sadeh
12:21 AM Revision 5912312c (ceph): rgw: remove some more globals from librgw
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:04 AM Revision 213a3f5e (ceph): rgw: fix identation
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:04 AM Revision d90298de (ceph): ceph-dencoder: don't use rgw types if configured without rgw
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh

03/13/2012

11:40 PM rgw Feature #2172 (Resolved): rgw: get chunks asynchronously
Chunks are read synchronously. We need to have a window of chunks that are read asynchronously (as with PUT). Yehuda Sadeh
11:38 PM rgw Feature #2171 (Rejected): rgw: asynchronously calculate md5
When doing a PUT we calculate the md5 of the content (used later for the etag) synchronously. We need to be able to c... Yehuda Sadeh
11:23 PM Revision a9d18975 (ceph): Merge branch 'master' of github.com:ceph/ceph
Johnw
11:22 PM Revision 60524aba (ceph): Added documentation for building the ceph documentation.
Johnw
09:59 PM Revision b9097619 (ceph): rgw: get rid of references to g_ceph_context where required
trickling down ceph context.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
08:48 PM Revision e6969258 (ceph): global: drop yellow warning on startup
Fixes: #2143
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
08:48 PM Revision e455d388 (ceph): doc: update project status/stability blurb
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Reviewed-by: Mark Kampe <mark.kampe@dreamhost.com>
Sage Weil
07:55 PM Revision e5934f10 (ceph): qa: kclient/file_layout.sh: ...
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
06:46 PM Revision 0a2068fc (ceph): Merge branch 'librados-cleanup'
Reviewed-by: Sage Weil <sage.weil@dreamhost.com> Josh Durgin
06:46 PM Revision 8f278647 (ceph): librados: split into separate files and remove unnecessary headers
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> Josh Durgin
06:46 PM Revision 5f92f338 (ceph): librados: move methods that require an IoCtx to IoCtxImpl
RadosClient still does a few different things, but at least it
no longer does all the work of an IoCtx.
Signed-off-b...
Josh Durgin
06:46 PM Revision db126279 (ceph): ObjectCacher: remove unused and crufty atomic sync operations
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> Josh Durgin
06:46 PM Revision 095c3a0e (ceph): OSDMap: make get_pools() const
Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com> Josh Durgin
06:46 PM Revision 16f99606 (ceph): osd_types: use uint64_t for ObjectExtent offsets and lengths
This is just client in-memory state, and allows us to address objects >4GiB,
to match the existing librados/Objecter ...
Josh Durgin
05:49 PM Revision b90354db (ceph): thrash: put client on separate machine from osds
This allows us to run kenrel clients (kclient, rbd) against the thrashing
cluster.
Sage Weil
05:09 PM Revision 5c9acbd8 (ceph): gitbuilder: put flavor last
in case we refine the field later Sage Weil
05:02 PM Revision 1a01ccaa (ceph): Pull from new gitbuilder.ceph.com locations.
Simplifies the flavor stuff into a tuple of
<package,type,flavor,dist,arch>
where package is ceph, kenrel, etc.
typ...
Sage Weil
01:56 PM Bug #2132: FAILED assert(!missing.is_missing(soid))
stxShadow saw this as well. Josh Durgin
01:45 PM Cleanup #2143 (Resolved): Remove ALL "don't use this product" warnings
Sage Weil
01:31 PM Feature #2145 (Resolved): doc gitbuilder
Sage Weil
12:28 PM Linux kernel client Bug #2064: ceph-client: messenger: nocrc flag not implemented correctly
Update: the commit had to be rebased, so it's id is now: 4d3e7aa992
Alex Elder
08:09 AM Linux kernel client Bug #2064: ceph-client: messenger: nocrc flag not implemented correctly
This is fixed by this commit:
086da4c6f8 libceph: fix inverted crc option logic
That is now present in the c...
Alex Elder
12:26 PM Linux kernel client Bug #2157: ceph: xattr: fix nanosecond display on i_rctime
This has been fixed in this commit:
260ac0e65b ceph: fix three bugs, two in ceph_vxattrcb_file_layout()
The comm...
Alex Elder
12:26 PM Linux kernel client Bug #2156: ceph: xattr: fix a possible buffer overrun bug
This has been fixed in this commit:
260ac0e65b ceph: fix three bugs, two in ceph_vxattrcb_file_layout()
The comm...
Alex Elder
12:26 PM Linux kernel client Bug #2155: ceph: xattr: wrong value assumed for "no preferred PG"
This has been fixed in this commit:
260ac0e65b ceph: fix three bugs, two in ceph_vxattrcb_file_layout()
The comm...
Alex Elder
11:01 AM rgw Bug #2170: librgw references g_ceph_context
Ouch. Mostly through dout, but there are other references. Yehuda Sadeh
10:40 AM rgw Bug #2170 (Resolved): librgw references g_ceph_context
2012-03-13T00:48:30.009 INFO:teuthology.task.workunit.client.0.err:OSError: /tmp/cephtest/binary/usr/local/lib/librgw... Sage Weil
09:31 AM rgw Feature #2169 (Resolved): rgw: api to control bucket placement
It'd be nice to be able to control which pool the bucket would be placed in when creating it. Yehuda Sadeh
12:02 AM Revision 98792e93 (ceph): rgw: add more meaningful tests instances of encoded objects
this completes #2140
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh

03/12/2012

11:46 PM Revision dfe50bcb (ceph): build-doc: use alternate virtualenv dir, if specified
The docs gitbuilder will use this to avoid rebuilding the virtualenv on
every build.
Signed-off-by: Sage Weil <sage@...
Sage Weil
10:22 PM Revision 096427d5 (ceph): remove dup ceph tasks from new thrash workloads
Sage Weil
10:12 PM Revision 6869c576 (ceph): test_idempotent: fix global_init call
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
10:01 PM Revision 401a5716 (ceph): qa: kclient/file_layout.sh poking
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
09:59 PM Revision 7e0e7ce1 (ceph): rgw: modify bucket instance for encoding test
This makes 'make check' happy, otherwise we need to create
a bucket name that starts with a period. This version is b...
Yehuda Sadeh
09:58 PM Revision 9969bba8 (ceph): FileStore: clarify debug/error output
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
09:58 PM Revision 6ca070ad (ceph): FileStore: ignore ERANGE and ENOENT on replay
The source object may either not exist or be the wrong size
during replay if the destination object was deleted in a ...
Samuel Just
09:58 PM Revision f20f6ea7 (ceph): test: add test_filestore_idempotent2
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
09:41 PM Revision e19417ef (ceph): Makefile.am, rgw: remove fcgi dependency where not needed
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
09:22 PM Revision 3ff5492c (ceph): rgw: add more meaningful test instances of some encoded objects
still need to add tests for other objects
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
08:22 PM Revision 31578c19 (ceph): rgw: tone down some log messages
dout(0) -> dout(1)
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
08:21 PM Revision 452b1248 (ceph): rgw: provide different default values for 'debug rgw'
Currently rgw and radosgw-admin require different chattiness
defaults.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdre...
Yehuda Sadeh
08:21 PM Revision 2c88f85c (ceph): config: alternative config options for global_init()
We want to be able to provide alternative default config values, than
the ones we set in common/config_opts.h. This c...
Yehuda Sadeh
08:13 PM Revision 3833ada8 (ceph): Made the example better with multiple roles.
Mark Nelson
07:33 PM Revision 0a61ffad (ceph): Added some example yaml files and an example parallel execution task.
Mark Nelson
07:11 PM Bug #2168 (Resolved): qa: plana won't select correct kernel for boot
Just need to run chef once before trying to install another kernel. Josh Durgin
05:38 PM Bug #2168: qa: plana won't select correct kernel for boot
Josh Durgin
04:26 PM Bug #2168 (Resolved): qa: plana won't select correct kernel for boot
the grub.cfg has default set to the kernel we want, but it boots up into a different one.
no console, so hard to d...
Sage Weil
07:08 PM Feature #2052 (In Progress): librbd: caching
Josh Durgin
07:08 PM Feature #2053 (New): librados: caching
Josh Durgin
07:01 PM Revision 968d29d1 (ceph): qa: use recent kernel for kernel_untar_build.sh
Happier on oneiric!
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
06:43 PM Revision 93164393 (ceph): rgw: switch ops log flag to use ceph config
It's turned on by default. So now we're using the
'rgw enable ops log' config param in ceph.conf, instead
of RGW_SHOU...
Yehuda Sadeh
06:22 PM Revision c9898f43 (ceph): filestore: fix op_num offset/labels
Start at 0, not 1.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
06:20 PM Revision 7c753498 (ceph): config: tmap to omap upgrade, true by default
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
05:00 PM Feature #2142 (In Progress): Add support for Ubuntu 12.04 (precise pangolin) in deb repository
Sage Weil
04:58 PM rgw Feature #2140 (Resolved): rgw: use new encode/decode macros
Yehuda Sadeh
03:01 PM rbd Feature #2167 (Duplicate): rbd: report the allocated size/space of a RBD with "rbd info <rbd name>"
With "rbd info <rbd name>" it reports the provisioned space, but not the allocated space (rbds are thin provisioned). Stefan Kleijkers
02:58 PM CephFS Bug #1737: ceph-fuse crash in xlist::remove
ubuntu@teuthology:/a/nightly_coverage_2012-03-12-b/727 Sage Weil
01:19 PM rgw Cleanup #2089 (Resolved): rgw: less dout(0) noise?
Fixed, commit:31578c19ffe9276a3e6495707e72180436ef3272. I've also set different 'debug rgw' default for rgw and rados... Yehuda Sadeh
01:09 PM Bug #2098 (Resolved): xfs/ext4 non-idempotent transaction
Sage Weil
12:18 PM Feature #2145 (In Progress): doc gitbuilder
Sage Weil
11:56 AM rgw Feature #2040 (Resolved): rgw: disable rgw log through ceph.conf
Sage Weil
10:07 AM Feature #2125: osd: put large xattrs in leveldb
Sage Weil
09:39 AM Feature #1956 (Resolved): rgw: revisit atomic GET/PUT
Sage Weil
04:50 AM Revision 2b9e7bc5 (ceph): clusters/fixed-3.yaml: 2 -> 6 osds
plana nodes have 3 scratch disks... use them! Sage Weil
04:32 AM Revision 51d817fe (ceph): Revert "disable s3tests on valgrind/lockdep until #2103 is fixed"
This reverts commit 9f757ca9511374f6565d74263e242c74e39f8a3f. Sage Weil
04:28 AM Revision af445189 (ceph): add rbd, kclient workloads to regression thrash collection
This will get us some kernel osd_client osd restart coverage. Sage Weil
04:16 AM Revision 3bba6b72 (ceph): Makefile: link libfcgi to librgw
Need this to make a linker error go away on my squeeze dev box. We
probably need to make sure librgw doesn't touch f...
Sage Weil
03:36 AM Revision 945218cc (ceph): qa/workunits/kclient/file_layout: escape *
Escape * so that it is expanded as root.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil

03/11/2012

10:24 PM Feature #2020 (Duplicate): collectd: submit plugin upstream
Sage Weil
09:47 PM Bug #2160 (Fix Under Review): active+recovering+degraded+backfill becomes active+clean+degraded w...
See wip-2160 for a fix. Sage Weil
09:41 PM rgw Cleanup #2166 (Resolved): rgw: make sure librgw doesn't link against libfcgi
Currently it does, and seems to need to to build on squeeze. Sage Weil
09:18 PM Bug #2165 (Resolved): osd: recovering ending with missing
... Sage Weil
09:16 PM Bug #2164 (Resolved): osd: scrub missing _, snapset attrs
... Sage Weil
08:03 PM Revision 71e6e62e (ceph): fix typo, ceph-fyuse -> ceph-fuse
Sage Weil
07:31 PM Revision a8497874 (ceph): filejournal: less log noise
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
03:20 PM Feature #2163: qa: full xfstests on rbd
iirc xfstests normally wants a block device it is free to mount/umount/reformat/etc. we can do this on top of rbd. Sage Weil
03:18 PM Feature #2163 (Resolved): qa: full xfstests on rbd
Sage Weil
02:49 PM Feature #1929 (Resolved): teuthology: log runtime
Sage Weil
04:01 AM Revision b84897e5 (ceph): use dbench workunit, not the autotest one
The autotest one uses an old tarball that doesn't build. Workunit assumes
the dbench package is installed.
Sage Weil
03:15 AM Revision 008cf7fd (ceph): autotest: pull from github.com/ceph/autotest
Sage Weil

03/10/2012

11:34 PM Revision 2124129e (ceph): workunit: include python2.7 path too
Sage Weil
04:42 PM Feature #2114 (Resolved): old sepia setup on new hardware
Sage Weil
04:42 PM Feature #2162 (Resolved): change gcov, notcmalloc gitbuilders to oneiric
Sage Weil
04:55 AM Revision 0851b747 (ceph): filestore: remove unused bool idempotent
This was from the old broken mechanism.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
01:07 AM Revision 6980ed98 (ceph): filestore: fix arguments
From a change that was rebased out; missed this caller.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
12:42 AM Revision ea34aa59 (ceph): Merge remote branch 'gh/wip-2098'
Reviewed-by: Samuel Just <samuel.just@dreamhost.com> Sage Weil
12:34 AM Revision 9b2a5c5d (ceph): filestore: sync object_map on _set_replay_guard()
We need to sync the object_map too. We can _almost_ check to see if there
are keys for the object and only do it the...
Sage Weil
12:32 AM Revision 51372a76 (ceph): filestore: guard replay of collection_add
- set guard on apply
- check guard on replay
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
12:32 AM Revision 6613b415 (ceph): filestore: guard collection_remove replay
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision 13afce83 (ceph): filestore: remove old post-idempotent transaction trigger_commit
The old strategy was to initiate a commit after any non-idempotent
transaction. This only worked if the transaction ...
Sage Weil
12:32 AM Revision c9550c59 (ceph): omap: add sync method to ObjectMap
Signed-off-by: Samuel Just <rexludorum@gmail.com> Sage Weil
12:32 AM Revision 808812a2 (ceph): filestore: use proper type for readdir_r tmp
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision a6077f6c (ceph): filestore: pass trans_num into _do_transaction
This gives us the <op_seq, trans_num, op_num> triple to identify every
constituent operation.
Signed-off-by: Sage We...
Sage Weil
12:32 AM Revision babb1e90 (ceph): add SequencerPosition type
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision e81bc971 (ceph): filestore: fgetxattr helpers/wrappers
Also, do the getxattr using fgetxattr, to avoid duplicating code. This is
slightly slower probably because we open a...
Sage Weil
12:32 AM Revision 3b47c6ba (ceph): filestore: maintain SequencerPosition during _do_transaction
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision f8a047f2 (ceph): filestore: implement _set_replay_guard, _check_replay_guard
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision e99940d1 (ceph): filestore: guard clone replay
- set guard xattr on clone, clone_range
- check before applying/replaying
Signed-off-by: Sage Weil <sage.weil@dreamh...
Sage Weil
12:32 AM Revision 63d2c93d (ceph): filestore: fix collection_rename error code
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
12:32 AM Revision dd284c4f (ceph): filestore: guard collection_rename replay
- check guard on replay
- set guard on apply
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
12:32 AM Revision 4ea6e0d2 (ceph): filestore: guard replay of basic collection ops
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil

03/09/2012

10:24 PM Revision 11844978 (ceph): levedb: fix commit
This got reverted back to the old commit, somehow.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
10:13 PM Revision 184b1802 (ceph): Merge branch 'master' of github.com:ceph/ceph
Sage Weil
10:12 PM Revision 7f9cfdd9 (ceph): RadosModel: fix omap_clear case in RemoveAttrsOp
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
10:03 PM Revision fce180df (ceph): Merge branch 'wip-rgw-encode'
Conflicts:
src/rgw/rgw_cls_api.h
Reviewed-by: Sage Weil <sage@newdream.net>
Sage Weil
10:01 PM Revision b66cdea6 (ceph): Makefile.am: update link dependencies for some unit tests
Yehuda Sadeh
10:01 PM Revision b4f05788 (ceph): test/encoding/import.sh: fix target directory
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@dreamhost.com> Yehuda Sadeh
10:01 PM Revision b71d814c (ceph): ceph-object-corpus: added 0.43
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@dreamhost.com> Yehuda Sadeh
10:01 PM Revision 799e6f97 (ceph): rgw: fix rgw_cls_list_ret ctor
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
10:01 PM Revision 09949b3d (ceph): rgw: various encoding related fixes
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
09:34 PM Revision e43546de (ceph): osd: fix watch_lock vs map_lock ordering
watch_lock is inside map_lock (and pg->lock), which means we need to
drop it to take pg->lock here. That means verif...
Sage Weil
08:26 PM Revision 80d92528 (ceph): osd: update_heartbeat_peers as needed
Before, we were being very careful about updating the heartbeat peers if
new PGs were created or when certain types o...
Sage Weil
05:05 PM Feature #2162 (Resolved): change gcov, notcmalloc gitbuilders to oneiric
Sage Weil
04:39 PM Bug #2098: xfs/ext4 non-idempotent transaction
merged. passing initial tests. will hammer on it more with sam's new thing. Sage Weil
03:31 PM Bug #2161 (Resolved): nonlinear scaling for PGMap::pg_stat encode
... Sage Weil
02:03 PM rgw Feature #2140: rgw: use new encode/decode macros
we should add some semi-meaningful test instances for these objects, otherwise all looks good Sage Weil
01:34 PM Bug #2103 (Resolved): osd: lockdep error on watch_lock
Sage Weil
11:07 AM Bug #2116: Repeated messages of "heartbeat_check: no heartbeat from"
Hrm... well, if that's the case then we should merge the wip-2116 as is, then... putting that on my list
Sage Weil
10:58 AM Bug #2116: Repeated messages of "heartbeat_check: no heartbeat from"
I'm not willing to say it's completely gone, but it is starting to seem like it.
In the past few days the cluster ...
Wido den Hollander
09:04 AM Bug #2160 (Resolved): active+recovering+degraded+backfill becomes active+clean+degraded when reco...
In a configuration with 3 replicas of each PG (I haven't tested with others), when one of the disks is replaces, some... Alexandre Oliva
08:16 AM CephFS Bug #2159 (Resolved): ceph-fuse: big_writes option not recognized
... Sage Weil
08:06 AM Revision 619244d8 (ceph): rgw: implement dump() for encoders
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:58 AM Revision 2277fb45 (ceph): rgw: add stubs for dencoder test
still need to add some content to the dump methods
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
12:15 AM Revision 71323ffe (ceph): Merge remote branch 'origin/wip-2139'
Conflicts:
src/cls_rgw.cc
src/rgw/rgw_rados.cc
src/rgw/rgw_rados.h
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@drea...
Yehuda Sadeh
12:14 AM Revision ddc1ab0c (ceph): rados.py: include setattr and rmattr
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just

03/08/2012

11:54 PM Revision 64e3ffb2 (ceph): Merge branch 'master' of ssh://github.com/ceph/ceph
Yehuda Sadeh
11:18 PM Revision ac28b603 (ceph): ceph: document the way files are laid out
This adds a document that I wrote about how Ceph client file data
is striped across Ceph objects to the repository. ...
Alex Elder
11:09 PM Revision 1ed1d466 (ceph): librados: fix unit test for omap_get_vals_by_key rename
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
11:06 PM Revision ce43bd53 (ceph): osd: format time nicely in ops_in_flight output
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
11:06 PM Revision 934ee487 (ceph): librados: fix map -> std::map in header, string -> std::string
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
11:02 PM Revision fafbddd9 (ceph): doc: example of diagnosing radosgw hang
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
10:55 PM Revision 99cb9945 (ceph): doc: explain how unfound objects happen
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
10:55 PM Revision 47ce7ade (ceph): doc: make osd failure example include >3 osds
More realistic.
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
10:46 PM Revision 80864a12 (ceph): testrados: fix omap_get_vals_by_keys call
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
10:30 PM Revision e7e3baeb (ceph): osd: add zero_to field to PG::OndiskLog; track zeroed region of pg log
Track which region of the log has been zeroed on disk. This may be
different from tail if 'osd preserved trimmed log...
Sage Weil
10:30 PM Revision ed6cc47b (ceph): filestore: use FL_ALLOC_PUNCH_HOLE to zero, when available
First try the FL_ALLOC_PUNCH_HOLE fallocate() flag. If we get EOPNOTSUPP,
fall back to writing zeros.
Check for fal...
Sage Weil
10:17 PM Bug #2103 (Fix Under Review): osd: lockdep error on watch_lock
Sage Weil
04:47 PM Bug #2103 (In Progress): osd: lockdep error on watch_lock
Sage Weil
10:16 PM Revision 1468d951 (ceph): osd: fix op_wq vs pg->lock ordering
map_lock
-> pg->lock
-> op_wq
Fixes: #2153
Signed-off-by: Sage Weil <sage@newdream.net>
Reviewed-by: Samuel Just...
Sage Weil
04:07 PM rgw Bug #2139 (Resolved): rgw: no cleanup of index update in case of cancelled operations
Fixed, merged at commit: 71323ffeda665defdd133d9fd3c40e08952082e4. Yehuda Sadeh
02:57 PM Feature #2158 (Duplicate): cephtool: helpful error/timeout when no monitor quorum
Sage Weil
02:44 PM Feature #975 (Resolved): osd: use fallocate(FL_PUNCH_HOLE) for zero
commit:e7e3baebd6d5d9e3dfbddc77a2312e9f8e4e9da8 Sage Weil
02:34 PM Feature #2088: msgr: refactor 2 threads to one
Maybe it's a good time to switch from poll to epoll (where available)? Yehuda Sadeh
02:30 PM Linux kernel client Bug #2157 (Resolved): ceph: xattr: fix nanosecond display on i_rctime
In ceph_vxattrcb_dir_rctime() the nanoseconds part of the time
displayed does not include leading 0's, which leads t...
Alex Elder
02:25 PM Bug #2153 (Resolved): osd: locking cycle on pg::_lock. op_tp::lock, OSD::map_lock
commit:1468d95101adfad44247016a1399aab6b86708d2 Sage Weil
09:27 AM Bug #2153 (Resolved): osd: locking cycle on pg::_lock. op_tp::lock, OSD::map_lock
... Sage Weil
02:10 PM Linux kernel client Bug #2156 (Resolved): ceph: xattr: fix a possible buffer overrun bug
In ceph_vxattrcb_file_layout(), if an inode has a preferred PG its
value is added to the formatted output buffer. I...
Alex Elder
01:56 PM rgw Cleanup #2089: rgw: less dout(0) noise?
sounds good to me! Sage Weil
01:40 PM rgw Cleanup #2089: rgw: less dout(0) noise?
debug rgw = 0 removes almost everything except two:
rgw_main.cc: dout(0) << "====== starting new request req=" <<...
Dan Mick
10:05 AM rgw Cleanup #2089: rgw: less dout(0) noise?
Actually, those are dout(10). The problem is that the default rgw log level is 20. Yehuda Sadeh
01:41 PM Linux kernel client Bug #2155 (Resolved): ceph: xattr: wrong value assumed for "no preferred PG"
In ceph_vxattrcb_file_layout() the layout for a file is printed.
The preferred placement group is extracted from t...
Alex Elder
01:14 PM Bug #2154 (Resolved): rados: bench seq should not segfault when blocksize doesn't match write blo...
I misunderstood the problem to begin with. The issue is when read size does not match the prior write size (I hadn't... Dan Mick
09:50 AM Feature #2149: osd: use omap for snap collections
Sage Weil
09:40 AM rgw Feature #1739 (Resolved): rgw: multipart upload should use manifest object
Sage Weil
08:56 AM Feature #1895 (Rejected): osd: detect duplicate requests by tracking per-client last_acked_tid in...
Sage Weil
08:54 AM Feature #1802: qa: test to exercise divergent osd logs
the backfill.py exercises a divergent backfill target. we just need to do the same on a non-backfill target. Sage Weil
07:46 AM rgw Feature #2066 (Resolved): rgw: make list_objects efficient
Done, merge in with rgw-omap changes at commit:89648a0d1abe06832f57db2ff8be85749f667094 Yehuda Sadeh
07:37 AM rgw Feature #1715 (Rejected): rgw: use RENAME osd operation to avoid slow CLONE operations
With new atomic scheme we're not using clone anymore (unless it's an old object for backward compatibility). Yehuda Sadeh
07:34 AM rgw Feature #2040: rgw: disable rgw log through ceph.conf
Yeah, this is the rgw user operations log, not the debug log. Yehuda Sadeh
07:25 AM rgw Feature #1847 (Resolved): rgw: revisit the way we store large objects
Revisited, solution was merged in commit:5f8bbf1ff2d547023c15628dd0e3a5711e584c2b (rgw-new-atomic). Yehuda Sadeh
07:23 AM rgw Feature #1872 (Resolved): rgw: only use shadow objects for large objects
This has long been done.
Pushing acls check to cls_rgw can be done later on, but that's a different issue.
Yehuda Sadeh
07:21 AM rgw Bug #2072 (Resolved): rgw: owner cannot change acl if it doesn't have bucket read permission
Fixed, merged in with swift-acls. Yehuda Sadeh
07:19 AM rgw Feature #2140 (In Progress): rgw: use new encode/decode macros
Yehuda Sadeh
06:58 AM Revision c0aa3319 (ceph): Merge branch 'master' of ssh://skinny/home/yehudasa/ceph
Yehuda Sadeh
06:53 AM Revision 5f8bbf1f (ceph): Merge branch 'wip-rgw-new-atomic'
Yehuda Sadeh
06:52 AM Revision 465a88ea (ceph): rgw: append the currect bucket marker when removing bucket
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
06:39 AM Revision 8ce3d229 (ceph): Merge branch 'master' of ssh://skinny/home/yehudasa/ceph
Yehuda Sadeh
06:35 AM Revision 89648a0d (ceph): Merge branch 'wip-rgw-omap'
Yehuda Sadeh
06:25 AM Revision 72d678b8 (ceph): cls_rgw: fix rgw_bucket_init_index
was failing to error in case header already existed
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
06:19 AM Revision eaafc578 (ceph): rgw: remove extra unused params from omap_get()
and also rename it to omap_get_all()
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
06:18 AM Revision 765962dc (ceph): rgw: add cls_cxx_map_clear
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
05:59 AM Revision afe2598a (ceph): leveldb: drop compaction unit test
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
01:20 AM Revision 4d6578ca (ceph): ReplicatedPG,librados: add filter_prefix to omap_get_vals
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
01:10 AM Revision e0744390 (ceph): rgw: use prefix filter for bucket listing
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
01:03 AM Revision 4c908442 (ceph): objclass, cls_rgw: add prefix to omap_get_vals()
Yehuda Sadeh
01:02 AM Revision a6b524f6 (ceph): librados: add higher level call for omap_get_keys() with prefix
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:46 AM Revision 7a72bf01 (ceph): Merge remote-tracking branch 'origin/wip_prefix' into wip-rgw-omap
Yehuda Sadeh
12:32 AM Bug #2146: Ceph -s crash and reports old ceph version
I have updated ceph package only. The binary used for getting statistics is in ceph-common. After upgrade it work fin... Maciej Galkiewicz
12:29 AM Revision ffcf62f6 (ceph): rgw: switch to new encoders/decoders
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
12:28 AM Revision ffae4dfe (ceph): enconding: add legacy decoder for 32 bit versions
In these cases if the version is smaller than the compat ver
we skip 3 bytes.
Signed-off-by: Yehuda Sadeh <yehuda@hq...
Yehuda Sadeh
12:05 AM Revision b0c9d58c (ceph): rbd: pass all mon addrs when mapping devices
Previously this repeated the address of the first monitor.
Fixes: #2152
Signed-off-by: Josh Durgin <josh.durgin@drea...
Josh Durgin

03/07/2012

10:07 PM Revision 4d0bcdc8 (ceph): objectstore: fix collection_move() encoding
This was broken in the original f43c3d958fe5c32ae647ffa715390ada51ae2650.
Signed-off-by: Sage Weil <sage.weil@dreamh...
Sage Weil
09:54 PM Revision 9dc7c653 (ceph): msgr: remove SimpleMessenger::get_ms_addr() in favor of Messenger::get_...
And fix the comments on set_ip.
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Reviewed-by: Sage Weil <sa...
Greg Farnum
09:54 PM Revision 65d170a8 (ceph): msgr: remove declaration of undefined SimpleMessenger::write_pid_file
Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Reviewed-by: Sage Weil <sage.weil@dreamhost.com>
Greg Farnum
09:21 PM Feature #2114 (In Progress): old sepia setup on new hardware
Sage Weil
09:20 PM Feature #2087 (In Progress): lightweight filestore workload generator
Sage Weil
09:20 PM Feature #2088 (In Progress): msgr: refactor 2 threads to one
Sage Weil
09:20 PM Feature #2053 (In Progress): librados: caching
Sage Weil
09:20 PM rgw Feature #2066 (In Progress): rgw: make list_objects efficient
Sage Weil
09:20 PM rgw Feature #1950 (In Progress): rgw: create S3/Swift ACL interoperability suite
Sage Weil
09:20 PM rgw Feature #830 (Resolved): rgw: swift per-object ACLs
Sage Weil
09:19 PM rgw Feature #2140 (Resolved): rgw: use new encode/decode macros
Sage Weil
09:08 PM Revision 0a59d08d (ceph): ReplicatedPG,librados: add filter_prefix to omap_get_vals
Signed-off-by: Samuel Just <rexludorum@gmail.com> Samuel Just
08:55 PM Revision 31762c00 (ceph): lock: Improved logging when there aren't enough nodes available to lock...
Mark Nelson
08:55 PM Revision 05a07dda (ceph): lock: Added a --locked flag to teuthology-lock.
Can be used to restrict searches based on lock status, e.g.
'teuthology-lock --list -a --locked false --status up' sh...
Mark Nelson
08:53 PM rgw Feature #2040: rgw: disable rgw log through ceph.conf
Dan Mick wrote:
> Yes please! :) rgw debug = 0 doesn't seem to have any effect. (you knew that.)
...but this i...
Dan Mick
08:47 PM rgw Feature #2040: rgw: disable rgw log through ceph.conf
Yes please! :) rgw debug = 0 doesn't seem to have any effect. (you knew that.) Dan Mick
08:34 PM Revision 6bf72019 (ceph): rgw: some minor cleanups
following a review
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
06:45 PM Revision 5f8ffdee (ceph): objclass: fix cls_cxx_map_write_header
Claiming the buffer instead of encoding it.
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@dreamhost.com>
Yehuda Sadeh
06:44 PM Revision 799c8aca (ceph): cls_rgw: fix debug message
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@dreamhost.com> Yehuda Sadeh
06:32 PM Revision 0b1037da (ceph): Merge remote-tracking branch 'gh/wip-doc'
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com> Sage Weil
04:56 PM Revision 9ba96c01 (ceph): osd: make degraded pgs count missing replicas as degraded objects
If a PG is smaller than it should be, make sure the missing replicas are
included in the degraded object count. This...
Sage Weil
04:14 PM Bug #2152 (Resolved): rbd: first monitor address is passed to kernel n times in rbd map
Sage Weil
02:45 PM Bug #2152 (Resolved): rbd: first monitor address is passed to kernel n times in rbd map
Using rbd map with mon0 (addr 10.10.10.4:6789) down, this happened:... Josh Durgin
02:41 PM Linux kernel client Bug #2081: msgr: spinlock badness?
This:
ubuntu@teuthology:/a/nightly_coverage_2012-02-20-b/12984
does not provide very much useful informatio...
Alex Elder
02:36 PM Feature #2127: Save kernel core dumps on all of our test machines
Researched this a bit this afternoon. I have used LKCD in the past and
find now it's fairly old. Here someone has ...
Alex Elder
02:11 PM Bug #2151 (Resolved): mon: pg creation retry is slow
right now if you are creating a pg and the osd crashes, the monitor will retry every once in a while, but that retry ... Sage Weil
02:00 PM Messengers Cleanup #2150 (Resolved): repair the Simple/Messenger interface
I made a start on this with my interface work, but there are still
1) a lot of public functions that ought to be pri...
Greg Farnum
10:51 AM Feature #2149 (Resolved): osd: use omap for snap collections
Sage Weil
09:55 AM Bug #2148 (Resolved): osd: class error return not propagated to client
Sage Weil
09:49 AM Feature #2147 (Resolved): objclass: add CLS_ERR macro
Otherwise, dumping errors through CLS_LOG hides them if debug objclass is not set. Yehuda Sadeh
09:06 AM Bug #2137 (Resolved): osd: refine degraded % calculation
commit:9ba96c0123c050f7e7c62ea644e240aed0c954fb Sage Weil
08:24 AM Bug #2146: Ceph -s crash and reports old ceph version
I'm curious how you managed to install new packages while keeping around an old binary, though. (Yes, that's definite... Greg Farnum
07:48 AM Bug #2146 (Rejected): Ceph -s crash and reports old ceph version
The problem is just that your ceph binary is out of date. You need to run v0.43 or later, due to the structure encod... Sage Weil
03:56 AM Bug #2146 (Rejected): Ceph -s crash and reports old ceph version
# ceph -s
terminate called after throwing an instance of 'ceph::buffer::malformed_input'
what(): buffer::malform...
Maciej Galkiewicz
05:03 AM Revision 184a9dda (ceph): mon: fix full osd detail
And use a helper to avoid dup code.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
05:03 AM Revision bd97923c (ceph): doc: fix misc typos, bad phrasing
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
05:03 AM Revision a300db39 (ceph): mon: assign severity to each health summary/detail item
These can be included in the detail dump in the future.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
01:17 AM Revision f266d5a2 (ceph): objclass, cls_rgw: update to use omap
Signed-off-by: Yehuda Sadeh <yehuda.sadeh@dreamhost.com> Yehuda Sadeh
01:05 AM Revision 2822fe50 (ceph): mon: include unfound count in health detail
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision 2b87d4f2 (ceph): osd: list might_have_unfound locations in query result
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision d72b8217 (ceph): doc: document some osd failure recovery scenarios
- simple osd failure
- ceph health [detail]
- peering failure ('down') state
- unfound objects
Signed-off-by: Sage W...
Sage Weil
01:05 AM Revision 2bec51a2 (ceph): doc: describe 'stuck' states we check for
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision 956e2e22 (ceph): mon: list nearfull/full osd detail
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision 75ad8979 (ceph): doc: diagnose full osd cluster
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision 25cceca0 (ceph): doc: slow osd requests
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision ab41d675 (ceph): doc: fix link
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision 0becb843 (ceph): doc: talk about mon failures a bit
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
01:05 AM Revision b0ed78c3 (ceph): doc: 2 words about radosgw failures
- restarting the daemon.
- using the admin socket
Signed-off-by: Sage Weil <sage@newdream.net>
Sage Weil
01:05 AM Revision 8b0bd127 (ceph): mon: refactor health, include optional detail
'ceph health' to get the usual summary, 'ceph health detail' to
additionally get a comprehensive list of problems fou...
Sage Weil
12:05 AM Revision db223363 (ceph): Merge branch 'wip-collmove'
Samuel Just

03/06/2012

11:48 PM Revision 85446298 (ceph): rgw: switch to omap api
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:15 PM Revision 6c0cb55e (ceph): leveldb: remove flawed unit test for now
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
10:53 PM Revision 7012c20b (ceph): librados: rename omap_get_vals_by_key to omap_get_vals_by_keys
merge fail
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
10:35 PM Bug #2137 (Fix Under Review): osd: refine degraded % calculation
wip-2137 Sage Weil
10:18 PM Revision 1fefb146 (ceph): librados: add high level omap calls
also rename get_vals_by_key to get_vals_by_keys
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
10:17 PM Revision f85ab71f (ceph): rgw: read bucket through tmap_get
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
10:17 PM Revision 7fc779a6 (ceph): rgw: fix warning
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
07:46 PM Revision 98f8219d (ceph): Merge branch 'wip_omap'
Reviewed-by: Sage Weil <sage.weil@dreamhost.com> Samuel Just
07:44 PM Revision d8dcb28e (ceph): librados: add tmap_put to ObjectWriteOperation
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision b85f7d7b (ceph): ReplicatedPG: add omap_header to recovery
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision d2bf68d1 (ceph): osdc: Add omap operation stubs to Objecter::ObjectOperation
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision 81c22dfb (ceph): librados: Added omap operations to librados
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision 82287986 (ceph): ReplicatedPG: Add omap ops to ReplicatedPG
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision 2abf3776 (ceph): RadosModel: Add omap operations to RadosModel
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision adace1cf (ceph): ReplicatedPG: transparently upgrade TMAP
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision 4c4fcea3 (ceph): osd: testing for tmap auto upgrade
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:44 PM Revision b6c2e839 (ceph): test_rados_api_aio: add omap
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> Samuel Just
07:03 PM Revision b52d4087 (ceph): Merge branch 'wip-1796'
Reviewed-by: Greg Farnum <gregory.farnum@dreamhost.com> Sage Weil
07:02 PM Revision 769ef369 (ceph): journaler: add generic write error handler
Specify a generic callback for any write error the journaler encounters.
This is more helpful than passing up write e...
Sage Weil
07:02 PM Revision 195301ef (ceph): mds: respawn when blacklisted
If we are blacklisted by the OSD cluster, it's because we were too slow
and were replaced by another ceph-mds. Respa...
Sage Weil
06:49 PM Revision 50682189 (ceph): Merge remote-tracking branch 'gh/wip-2105'
Reviewed-by: Samuel Just <samuel.just@dreamhost.com> Sage Weil
06:24 PM Revision 86186405 (ceph): .gitignore: src/ocf/rbd
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
06:23 PM Linux kernel client Bug #2064 (In Progress): ceph-client: messenger: nocrc flag not implemented correctly
I've been unable to test rigorously today so I analyzed the affected code.
If the "nocrc" flag is supplied, the ce...
Alex Elder
03:22 PM Linux kernel client Bug #2064: ceph-client: messenger: nocrc flag not implemented correctly
The problem also exists in write_partial_msg_pages(). In other words, it
looks like neither the reader nor the writ...
Alex Elder
05:46 AM Linux kernel client Bug #2064: ceph-client: messenger: nocrc flag not implemented correctly
This is a trivial fix. The only thing that needs to be done really is
testing the result to make sure things don't ...
Alex Elder
05:34 PM Revision 2a18c3e1 (ceph): nuke: unmount osd data directories
This helps us avoid reboot to clean up osd data directories that are left
mounted.
Sage Weil
05:19 PM Revision e3b4ba99 (ceph): filestore: create snap_0 on mkfs
If we create a new filestore, apply one transaction, and then crash, we
want to make sure roll back to a consistent r...
Sage Weil
05:19 PM Revision a14d44fc (ceph): filestore: drop useless read_op_seq() arg
Signed-off-by: Sage Weil <sage@newdream.net> Sage Weil
05:14 PM Revision b78b725d (ceph): Merge pull request #9 from fghaas/ocf-ra
OCF resource agents: add rbd
Reviewed-by: Sage Weil <sage@newdream.net>
Reviewed-by: Tommi Virtanen <tommi.virtanen@...
Sage Weil
04:40 PM Feature #2145 (Resolved): doc gitbuilder
it can automatically rsync to ceph.newdream.net/docs/<branch or tag name> Sage Weil
03:54 PM Bug #2136 (Resolved): leveldb fails unit tests in gitbuilders
Sage Weil
03:45 PM Subtask #1919 (Resolved): implement the key value interface in terms of leveldb
Samuel Just
03:45 PM Subtask #1920 (Resolved): Update OSD to use ObjectStore tmap implentation (object recovery must c...
Samuel Just
03:45 PM Subtask #1917 (Resolved): create interface for backing key-value store and create object map impl...
Samuel Just
03:45 PM Subtask #1918 (Resolved): create mock key-value store and tests for the related object map implen...
Samuel Just
03:45 PM Subtask #1915 (Resolved): Create trivial implementation for the object map interface (using curre...
Samuel Just
03:45 PM Subtask #1916 (Resolved): add tests ObjectStore level tests for tmap operations
Samuel Just
03:45 PM Subtask #1914 (Resolved): Create interface for object map implementation
Samuel Just
03:45 PM Subtask #1913 (Resolved): Add tmap operations to ObjectStore interface
Samuel Just
03:43 PM Feature #1571 (Resolved): osd: non-trivial map object
Samuel Just
03:42 PM Feature #1571 (Closed): osd: non-trivial map object
Samuel Just
02:20 PM CephFS Feature #2144 (Resolved): mon: improve mds health checks
e.g. replay, rejoin, resolve, etc. should be warning, and so forth. Sage Weil
11:27 AM Cleanup #2143: Remove ALL "don't use this product" warnings
should we leave it on ceph-mds startup? Sage Weil
11:26 AM Cleanup #2143 (Fix Under Review): Remove ALL "don't use this product" warnings
wip-2143 Sage Weil
11:21 AM Cleanup #2143 (Resolved): Remove ALL "don't use this product" warnings
We are about to push this product out to Open Stack and Canonical, and will be charging for support.
"Do not use thi...
Anonymous
11:10 AM CephFS Bug #1796 (Resolved): mds: exit cleanly on EBLACKLISTED
Sage Weil
10:57 AM Bug #2105 (Resolved): filestore: mkfs does not create initial snap
Fix is merged.. this'll be in v0.44.
Thanks for the bug report (and detailed log!)
Sage Weil
09:33 AM Bug #2105 (Fix Under Review): filestore: mkfs does not create initial snap
Sage Weil
10:21 AM CephFS Bug #1318: directories disappear across multiple rsyncs
I've just run into this on 0.43 using the ceph kernel module in 3.2.7. My symptoms are that a repeated rsync from a ... Jonathan Dieter
09:34 AM rgw Bug #2139 (Fix Under Review): rgw: no cleanup of index update in case of cancelled operations
Sage Weil
09:26 AM Feature #1655: gitbuilder aggregator page
We've been using this script for the past month or so, and it's been pretty handy. Here's an ajax version of the abov... Jimmy Tang
08:58 AM Revision affda7c0 (ceph): rbd OCF RA: fix whitespace inconsistency
Signed-off-by: Florian Haas <florian@hastexo.com> Florian Haas
08:25 AM Bug #2116: Repeated messages of "heartbeat_check: no heartbeat from"
I ran multiple I/O loads on the cluster last night and performed multiple tests, but this problem didn't come back ye... Wido den Hollander
07:17 AM Feature #2142: Add support for Ubuntu 12.04 (precise pangolin) in deb repository
I'll do this for v0.44. Sage Weil
01:25 AM Feature #2142: Add support for Ubuntu 12.04 (precise pangolin) in deb repository
This add is quite usefull for me, because I use chef (I will try to improve the current cookbooks) and chef use "node... Guilhem Lettron
01:21 AM Feature #2142 (Resolved): Add support for Ubuntu 12.04 (precise pangolin) in deb repository
Precise Pangolin is in beta now, it can't be used with the current repository. Guilhem Lettron
06:48 AM Revision d9d5cf2e (ceph): Merge remote branch 'gh/wip-msgr-interface'
Reviewed-by: Sage Weil <sage@newdream.net> Sage Weil
05:42 AM Revision d36b6a90 (ceph): osd: use new collection_move() operation
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
05:42 AM Revision 14929e4d (ceph): filestore: implement OP_COLL_MOVE
Equivalent to OP_COLL_ADD, OP_COLL_REMOVE.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
05:41 AM Revision f43c3d95 (ceph): objectstore: OP_COLL_MOVE
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
05:41 AM Revision 5137f64f (ceph): objectstore: use enum for OP_*
Enforce no dups.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
05:15 AM Revision d8740094 (ceph): objectstore: remove _fake_writes, _get_frag_stat
Also only implemented by ebofs.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
Sage Weil
05:12 AM Revision 737132be (ceph): filestore: drop trim_from_cache, is_cached
These were used for read optimizations in ebofs; I don't think they'll
come back.
Signed-off-by: Sage Weil <sage.wei...
Sage Weil
05:09 AM Revision 3cd5e31a (ceph): objectstore: remove cruft
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
05:09 AM Revision 97993f6c (ceph): filestore: remove collection, attr faking
Useless functionality from the dark ages of development, when xattrs were
scarce.
Signed-off-by: Sage Weil <sage.wei...
Sage Weil
01:28 AM Revision d0c6ad0f (ceph): rgw: cancel bucket update if operation failed
Beforehand the operation was leaking.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
01:27 AM Revision 14857582 (ceph): cls_rgw: add cancel operation
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh

03/05/2012

11:28 PM Revision 0661ea1d (ceph): Merge branch 'master' of ssh://github.com/ceph/ceph
Yehuda Sadeh
11:00 PM Revision 4e9ebd6b (ceph): rgw: make sure correct locator is used
Or more correct: locator is not used where not needed.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
11:00 PM Revision e0650ce3 (ceph): rgw: implement copy using new scheme
for some reason target tail uses locator, this needs to be
fixed.
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
11:00 PM Revision 53a2ceb7 (ceph): rgw: don't use locator for multipart uploads
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:00 PM Revision dfcb8412 (ceph): rgw: multipart object working with manifest
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:00 PM Revision 578c0510 (ceph): rgw: manifest object contains source offset info
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:00 PM Revision d409c313 (ceph): rgw: basic functionality of new atomic get/put works
get/put of objects works. Stuff that is known to be broken:
copy object
Also, going through the code, we can probabl...
Yehuda Sadeh
11:00 PM Revision 20244d64 (ceph): rgw: get_obj uses manifest
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net> Yehuda Sadeh
11:00 PM Revision c076e351 (ceph): rgw: atomic objects hold manifest header
When writing an object we update where all the chunks of this object
reside.
Signed-off-by: Yehuda Sadeh <yehuda@hq....
Yehuda Sadeh
11:00 PM Revision e36940af (ceph): rgw: atomic processor writes to shadow object
And the first chunk is going to the head object in the end
Signed-off-by: Yehuda Sadeh <yehuda@hq.newdream.net>
Yehuda Sadeh
10:51 PM Bug #2126 (Duplicate): osd: recover_primary did nothing when num_missing==1
Sage Weil
10:45 PM Bug #2105: filestore: mkfs does not create initial snap
Thanks for the report!
The mkfs doesn't create an initial snap, so if we crash/stop before creating one, our first...
Sage Weil
10:35 PM CephFS Bug #1796 (Fix Under Review): mds: exit cleanly on EBLACKLISTED
Sage Weil
10:35 PM Revision ed0f6053 (ceph): Merge remote branch 'gh/wip-swift-acls'
Lightly-reviewed-by: Sage Weil <sage@newdream.net> Sage Weil
10:21 PM Revision 3e95dfdf (ceph): osd: delay non-replayed ops during replay
If we get new (non-replayed) ops during replay, those need to wait until
after the replayed ops are ordered and appli...
Sage Weil
10:21 PM Revision 702f09ea (ceph): librados: close narrow shutdown race
timer.shutdown() will drop and retake the lock, so set DISCONNECTED first
to avoid a message slipping in and reaching...
Sage Weil
10:21 PM Revision 743da9bd (ceph): osd: don't trust pusher's data_complete
The pusher doesn't know what clone_overlap we'll see, so it has no idea
if we are data_complete from our perspective,...
Sage Weil
10:20 PM Revision e1a9e18b (ceph): osd: warn if recovery still has missing at end
We shouldn't get to this point. If we do, recover_primary didn't do what
it needed to. Dump the remaining missing s...
Sage Weil
10:14 PM Bug #2098 (In Progress): xfs/ext4 non-idempotent transaction
Sage Weil
09:34 PM Revision 14936747 (ceph): Use non-zero exit status if any tests failed
Fixes: #1989 Josh Durgin
08:30 PM Revision c31b8696 (ceph): OCF resource agents: add rbd
Add a resource agent for mapping, unmapping and monitoring RBD devices.
Maps an RBD on start, unmaps it on stop. Che...
Florian Haas
03:53 PM Feature #2141 (Resolved): ceph: 'object map <poolname> <objectname>' or similar
'ceph pg map 1.2' will tell you what osd handles a pg... having the same thing that lets you provide pool and object ... Sage Weil
02:32 PM rgw Feature #2140 (Resolved): rgw: use new encode/decode macros
so we can safely revise encoding for the structures in a forward/backward compatible way Sage Weil
02:18 PM Bug #2135 (Resolved): cephtool: osdc/Objecter.cc: 375: FAILED assert(initialized)
Sage Weil
02:17 PM Bug #2133 (Resolved): osd: recovery_complete
Sage Weil
02:17 PM Bug #2082 (Resolved): osd: broken queuing during replay
Sage Weil
01:23 PM Bug #2116: Repeated messages of "heartbeat_check: no heartbeat from"
I tried to compile the wip-2116 and master branches today, that failed. So I haven't been able to test the new code y... Wido den Hollander
12:39 PM rgw Bug #2120 (Resolved): rgw: atomic write guard doesn't scale well
Yehuda Sadeh
12:37 PM rgw Bug #2120: rgw: atomic write guard doesn't scale well
A fix for that was already pushed (backing off when identifying a race in atomic write scheme). However, a more compl... Yehuda Sadeh
11:43 AM Bug #1900 (Resolved): Fix detection and build issues with libcrypto++
commit:c9e028f4e5287ef6b7b237fca9ee34aa78ae1212 Sage Weil
10:51 AM Linux kernel client Bug #1600 (Can't reproduce): general protection fault during blogbench
Sage Weil
10:51 AM Linux kernel client Bug #1615 (Can't reproduce): null pointer dereference in ceph_msg_new
Sage Weil
10:50 AM Linux kernel client Bug #1597 (Can't reproduce): general protection fault during mount
Sage Weil
10:50 AM Linux kernel client Bug #2069 (Need More Info): client crash during kernel_untar_build rm -r step
Sage Weil
10:48 AM Linux kernel client Bug #1793 (Need More Info): NULL pointer dereference at try_write+0x627/0x1060
Sage Weil
10:45 AM rgw Bug #2063 (Resolved): rgw: access key shouldn't contain chars that need to be url encoded
Sage Weil
10:43 AM Feature #2087: lightweight filestore workload generator
*task Samuel Just
10:42 AM Feature #2087: lightweight filestore workload generator
Mark suggested that this would be a good talk for Joao. Samuel Just
10:42 AM Bug #1989 (Resolved): teuthology: error in ceph.log didn't make teutholgy return error code
Josh Durgin
10:40 AM CephFS Bug #2071 (Need More Info): kclient: pjd mkfifo failures
Sage Weil
10:40 AM rgw Bug #2035 (Resolved): rgw: bucket removal fails
Sage Weil
10:39 AM Bug #2042 (Need More Info): mon: crash in LogMonitor::update_from_paxos
Sage Weil
10:37 AM Feature #2003: limit XFS extent fragmentation for rbd
librbd caching will help a lot! Sage Weil
10:35 AM rgw Bug #1948 (Resolved): rgw: need to read intent log in chunks
Sage Weil
10:33 AM rgw Bug #2027 (Need More Info): rgw -> apache miscommunication
Sage Weil
10:08 AM rgw Bug #1864 (Resolved): rgw: atomic bucket info
Sage Weil
10:07 AM Bug #1873 (Won't Fix): crush_rule type is inconsistent
Sage Weil
10:06 AM CephFS Bug #1811 (Need More Info): 2 pjd chown tests failed on cfuse
Sage Weil
10:02 AM Bug #1627 (Need More Info): ceph-mon memleak if ceph-osd cluster ip is not reachable, but public ...
Sage Weil
10:01 AM CephFS Bug #1047: mds: crash on anchor table query
Sage Weil
09:57 AM CephFS Bug #1682 (Resolved): mds: segfault in CInode::authority
calling this resolved too. Sage Weil
09:51 AM CephFS Bug #1549 (Resolved): mds: zeroed root CDir* vtable in scatter_writebehind_finish
using clean shutdown now, yay Sage Weil
09:50 AM rgw Bug #1822 (Closed): radosgw can be slow to respond to requests
several other bugs are open to track specific problems Sage Weil
09:49 AM rgw Bug #2139 (Resolved): rgw: no cleanup of index update in case of cancelled operations
Yehuda Sadeh
09:33 AM Bug #2128 (Rejected): filestore: check() fails during sync
actually, i hit this running the new code with the filestore transaction leak bug fix reverted (to reproduce the btrf... Sage Weil
07:49 AM Linux kernel client Tasks #2138 (Resolved): rbd: run xfstests on a local XFS filesystem over RBD
This still can't really be done cleanly in a teuthology script, but
I would like to run xfstests on an RBD client sy...
Alex Elder
07:23 AM Bug #2137 (Resolved): osd: refine degraded % calculation
based on the recent pg state updates... Sage Weil

03/04/2012

01:40 PM Bug #2136 (Resolved): leveldb fails unit tests in gitbuilders
always ... Sage Weil
01:36 PM Bug #2082: osd: broken queuing during replay
Hit this again and dug into the logs. It's a bug on the osd... testrados is using the data, which has replay enabled... Sage Weil
05:01 AM Revision 75cbed61 (ceph): DBObjectMap: remove stray ;
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> Sage Weil
 

Also available in: Atom