Activity
From 05/04/2011 to 06/02/2011
06/02/2011
- 11:59 PM Revision 637dfc3e (ceph): rados_sync: add test for temp file deletion, fix
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:58 PM Revision d4edd17c (ceph): rgw: multipart: use locator on created parts
- 11:48 PM Revision 0f3224e1 (ceph): rados_sync: in export, download, then rename
- Download files to a 'temporary' name and then rename them when they are
complete. If the download gets aborted halfwa... - 10:24 PM Revision 37666185 (ceph): rgw: multipart additions and fixes
- 10:14 PM Revision 6fd694c3 (ceph): Remove unneeded libcrush1 files
- 10:13 PM Revision d6bbf3e5 (ceph): mds: journal parents of srci when srcdn is remote
- If srcdn is a remote dentry, we will be journaling the src inode to update
the mtime, but we need to ensure the paren... - 10:04 PM Revision ce5f0e71 (ceph): Move interactive and cfuse into tasks.
- 09:51 PM Revision 806646b0 (ceph): journaler: also initialize safe_pos
- on reread_head. Keep consistent across the two methods.
Signed-off-by: Sage Weil <sage@newdream.net> - 09:50 PM Revision a13b6643 (ceph): journaler: fix trim crash after standby-replay -> active
- The reread_head method needs to initialize trimming_pos (like read_head
does) or else we get confused later.
Signed-... - 09:14 PM Revision 7ca240bf (ceph): mds: cleanup rename_prepare a bit
- Use *srci tmp.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:57 PM Revision 0bcd9ac7 (ceph): vstart.sh: turn down debug ms
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:45 PM Revision 4d03e159 (ceph): rgw: some more multipard upload list
- 08:33 PM Revision 52bf3fce (ceph): rgw: extend multipart list parts response
- 08:24 PM Revision a670b4b3 (ceph): osd: implement clonerange
- Clone ranges of bytes between objects, provided
- src object locators match dest object
- src objects are not miss... - 08:24 PM Revision fc4cc399 (ceph): osd: give obc refs to RepGather
- Just give the ref to RepGather instead of doing a get and put.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 08:16 PM Revision 4cf342a1 (ceph): mds: pin inode while one renamed_files list
- Make sure the inode is pinned while it is on the LogSegment::renamed_files
list. Avoids a crash when an inode on tha... - 07:25 PM Bug #1098 (Closed): mds never coming "up:active" awaits in "up:creating"
- 07:11 PM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- Okay, there was an issue with the behavior with the MDS locks (they didn't do what I thought they did). I added a ne...
- 06:59 PM Linux kernel client Bug #1136 (Resolved): mempool_destroy failure on umount
- on current master, i was doing umount on a sluggish cluster, and...
- 06:23 PM CephFS Bug #1117: mds: rename rollback broken on slaves during replay
- It seems to also be broken on the master -- I've been testing cross-MDS rename ops and of course you see a lot of rol...
- 06:15 PM Bug #1118: Crash OSD after upgrdae from 0.28.1 to 0.28.2
- This looks like the same problem as #1127. If you're not using ext3/4 or btrfs, it almost certainly is. Can you try t...
- 06:11 PM Bug #1127: RBD got silent after 1 month
- Hi Yoshi, I put a fix in the filestore_debugging branch, which will generate a new package in about 1/2 an hour.
L... - 05:39 PM Revision b152a93c (ceph): rgw: more cleanup
- 05:28 PM Revision 3546cfdd (ceph): rgw: some cleanup
- 04:49 PM Revision 50731646 (ceph): rgw: multipart upload parser test util
- 04:39 PM Revision 2f3f36ab (ceph): rgw: fix multipart upload complete parser
- 04:30 PM Revision 711a77cf (ceph): rgw: multipart complete fix
- 04:19 PM Revision 0cce0a5e (ceph): filestore: allow clone_range to different offsets
- The old OP_CLONERANGE would only clone a range of bytes at the same offset
in both objects. Add an OP_CLONERANGE2 op... - 04:17 PM Revision 502baeab (ceph): filestore: fix fallback/slow do_clone_range
- We need to seek to the appropriate offsets on the src and destination
fd's for this to do the right thing.
Signed-of... - 04:17 PM Revision 6ca168ed (ceph): filestore: fix fallback/slow do_clone_range
- We need to seek to the appropriate offsets on the src and destination
fd's for this to do the right thing.
Signed-of... - 04:09 PM Revision 95163e94 (ceph): Fetch ceph binary tarball independently on every node.
- Avoids shuffling the bytes through the controlling node.
Use sha1 file to make sure everyone gets the same version. - 03:58 PM Cleanup #1135 (Resolved): d_type cleanup
- the codebase seems to have lots of this going on:
src/mds/CDir.cc:726: if (dn->get_linkage()->get_remote_d_ty... - 03:21 PM Bug #1134: rados export --delete-after can't clean up after a crash
- I guess I should add that manually removing that file from the exported directory makes it work again!
Also, it wo... - 03:16 PM Bug #1134 (Resolved): rados export --delete-after can't clean up after a crash
- I was using rados export to dump out a pool and it was taking a long time so I ctrl-c'd it. Now when I do it on that ...
- 03:13 PM RADOS Bug #1129 (Won't Fix): sort out libcrush
- meh, let's not worry about it until someone needs libcrush.so.
- 03:11 PM CephFS Bug #1132 (Resolved): mds: missing parent in rename metablob
- commit:d6bbf3e5fbe1df26d1bfe6f695ca52cfbb3694b2
- 01:19 PM CephFS Bug #1132 (Resolved): mds: missing parent in rename metablob
- single mds, fsstress -p 30 workload...
- 03:07 PM CephFS Bug #1133 (Resolved): mds: journaler failed assertion on standby-replay -> replay
- fixed by commit:a13b66436561bfe86f4907d18d2ea7762632d36d
- 02:04 PM CephFS Bug #1133 (Resolved): mds: journaler failed assertion on standby-replay -> replay
- fsstress workload. kill master mds. standby crashes with:...
- 12:29 PM Bug #1131 (Resolved): OSD assert failure in update_heartbeat_peers()
- Probably fixed in current stable: c5470e0f855b246cfbde6982ca90f565e7074600. Let us know if it persists!
- 12:20 PM Bug #1131 (Resolved): OSD assert failure in update_heartbeat_peers()
- I'm not sure I can reproduce it, because my system state is a bit out of whack due to a previous bug (#1130), but I'v...
- 09:54 AM Linux kernel client Bug #1096 (Resolved): LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Thanks Jeff!
- 02:51 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Hi , I apply the patch ,to verify this bug,
run "./fsstress -d /mnt/ceph/fstest -l 1 -n 10000 -p 1 -v" , pass.
r... - 04:20 AM Revision 7e2e4779 (ceph): mon: make sure osd paxos is writeable before doing timeouts
- The osd paxos machine has to be writeable before we can update it.
Fixes: #1130
Signed-off-by: Sage Weil <sage.weil@... - 12:05 AM Revision c5470e0f (ceph): OSD: don't keep old connection over new one in update_heartbeat_peers
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
06/01/2011
- 11:34 PM Revision 780322db (ceph): boto_tool: add get_bucket_acl
- Signed-off-by: "Colin McCabe" <colin.mcccabe@dreamhost.com>
- 11:28 PM Revision e11958b2 (ceph): Merge branch 'stable' into next
- 11:23 PM Revision 59501e1d (ceph): Merge branch 'stable'
- 11:04 PM Revision de0f0c72 (ceph): Refactor to use Cluster and Remote, to evaluate the new APIs.
- 10:13 PM Revision 65dc8411 (ceph): rgw: implement list multipart
- still partially implemented
- 09:54 PM Bug #1130 (Resolved): monitor crash in PaxosService:propose_pending()
- This should be fixed by commit:7e2e4779e4323429167af36e9a5fb9741c075e96. Thanks for the report!
- 04:36 PM Bug #1130 (Resolved): monitor crash in PaxosService:propose_pending()
- While doing some failure testing, one of the ceph monitors crashed. I have 6 osds, 3 monitors, and 3 mds servers run...
- 06:42 PM Revision e340bfe1 (ceph): dout: use recursive mutex for dout
- Using a recursive mutex for dout is desirable because it allows us to
survive situations like this:
> foo() { dout <... - 06:21 PM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
- Hi ,
ceph_set_page_dirty still exist igrab ,
i merge the patch to ceph-cleint-standalone,
run fsstress, still hit... - 06:18 PM Revision 44770df8 (ceph): lockdep: fix shadowed global, add printout
- Fix a bug that was keeping lockdep from starting. Add a printout that
lets the user know that lockdep is enabled.
Si... - 04:44 PM Revision 9b37f4fa (ceph): Allow embedded '\0' in bufferlists when copying to std::string.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 04:36 PM RADOS Bug #1129 (Won't Fix): sort out libcrush
- librados and libceph now statically link in crush code. Should it be a .so? Should be provide a .so anyway, for thi...
- 04:16 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- This should be fixed by commit:85defe76f7e2a0b3d285a3be72fcffce96629b5c, pushed to the master branch. Can you test an...
- 11:35 AM Linux kernel client Bug #1096 (In Progress): LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Scratch that, something a bit more subtle is going on.
- 11:14 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- The problem is a short O_DIRECT read that hits EOF. This seems to fix it for me:...
- 12:19 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- git ceph-client commit 98cc99822dac96710a8b64bdc2be4eccffc78956 ,
hand compiling , btrfs+ ubuntu 10.10+2.6.39+ ..
... - 04:09 PM CephFS Bug #1128: clustered mds: failed verify_subtree_bounds
- Oh right. Logs and core dump in:
kai:~gregf/logs/fstress/replay_bad_bounds - 04:04 PM CephFS Bug #1128 (Resolved): clustered mds: failed verify_subtree_bounds
- ...
- 09:14 AM Bug #1127: RBD got silent after 1 month
- Yoshi, can you attach 'ceph osd dump -o - 26', 'ceph osd dump -o -', and 'ceph pg dump -o -' outputs?
- 08:43 AM Bug #1127: RBD got silent after 1 month
- Looked into this a bit on irc yesterday. This part of the osd log looks problematic - there's only one osd, so the pg...
- 12:03 AM Revision 7c6c6a9e (ceph): rados_sync: don't hash paths with periods
- A period is not such a bad character.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
05/31/2011
- 11:57 PM Revision 4870393a (ceph): test_rados_tool.sh: test hashed paths
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:45 PM Revision b4bc1c68 (ceph): rados export: better name mangling rules, fix test
- Introduce a versioning scheme for name mangling, so that we can change
it in the future if we want to.
For names tha... - 11:05 PM Revision 5dd0e122 (ceph): rgw: handle multipart completion
- still wip
- 10:32 PM Revision d29b3b77 (ceph): rgw: parser for multi upload completion
- 10:01 PM Revision 7a474b10 (ceph): Use orchesta.remote as a higher-level wrapper, stop worrying about host...
- This changes just first caller in a series of many; the rest will change
once a role-based API is in place. - 10:01 PM Revision 33c39ab5 (ceph): rados_sync: prefix user extended attributes
- Start user extended attributes with USER_XATTR_PREFIX.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 09:59 PM Revision 0806e65b (ceph): rgw: some more xml reshuffling
- 09:59 PM Revision 9970b86c (ceph): Wrap Remote._runner in staticmethod() or it gets mistaken for a method.
- It used to get an extra self argument, and mistook that as client.
- 09:33 PM Revision dc9aaacf (ceph): Add a pretty wrapper on top of Paramiko and run.run.
- Most importantly right now, it knows its name, and can
prettyprint it. - 09:31 PM Revision f5d6be6e (ceph): rgw: move generic xml parsing code to some shared location
- 09:31 PM Revision 5875f796 (ceph): Remove dead code.
- 09:28 PM Revision efee7466 (ceph): objecter, osd: clonerange operation
- Add a src_oids field to MOSDOp, referenced by a new CLONERANGE osd op type
that will clone data from one object to an... - 08:58 PM Revision 07c1989a (ceph): librados: implement aio_flush
- Implement a per-ioctx flush that blocks until all previously submitted
aio operations on the ioctx are safe. Each ai... - 08:51 PM Revision 7d4bb120 (ceph): Initial import.
- Currently hardcoded to run dbench, not modular, and the remote
execution API is clumsy. - 08:46 PM Revision 6db2a4e2 (ceph): crushtool: error out if uniform weights vary
- Fixes: #1075
Signed-off-by: Sage Weil <sage@newdream.net> - 08:28 PM Revision 35b19a41 (ceph): osd: fix ScrubFinalizeWQ::_clear condition
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 07:58 PM Revision 1528d2c4 (ceph): debian: depend on libboost-dev >= 1.34
- for statechart. Partially fixes #1124.
Signed-off-by: Sage Weil <sage@newdream.net> - 05:45 PM Bug #1127 (Resolved): RBD got silent after 1 month
- RBD got silent after about 1 month running.
Although I restarted the daemons, the symptom doesn't go away.
Attached... - 04:37 PM Revision 0cfa911f (ceph): osd: don't leak Connection reference
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:37 PM Revision 8aa67aa4 (ceph): osd: ignore old/stale heartbeat messages
- If we get heartbeat messages from old epochs from peers that are not
current, drop them and mark the connection down.... - 04:37 PM Revision e5c9100b (ceph): osd: fix map sharing due to heartbeats
- - share the map with the cluster addr
- use the new {note,get}_peer_epoch helpers to do it sanely
- don't share if we... - 02:26 PM RADOS Feature #1126 (Rejected): crush: extend rule definition
- The current rule command structure does not allow you to do something like:
- pick 2 racks
- pick 2 devices under... - 01:57 PM Feature #511 (Resolved): librados: implement flush
- 01:44 PM RADOS Feature #1075 (Resolved): crushtool: warn if uniform item weights vary
- 01:15 PM rgw Subtask #1125 (Resolved): osd: support for merging/cloning several objects into one final object
- 12:34 PM Bug #1124: Depend on new enough Boost
- and
3. ceph.spec.in - 11:47 AM Bug #1124 (Resolved): Depend on new enough Boost
- Ensure that we depend on a new enough libboost to build successfully. Do this in
1. debian/control
2. autoconf
... - 11:34 AM Feature #1123 (Resolved): qa: small but completely functional suite
- 10:57 AM Bug #906 (Can't reproduce): clustered mds: lchown not setting uid/gid
- 10:53 AM CephFS Bug #1111 (Resolved): file lock requests in wait queue not getting cleaned up after process exit
05/30/2011
- 09:45 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
echo 'file fs/ceph/caps.c +p' > /sys/kernel/debug/dynamic_debug/control
logs attached .- 08:29 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
Hi ,i git ceph-client master branch:
commit 98cc99822dac96710a8b64bdc2be4eccffc78956
Author: Sage Weil <sage@ne...- 09:14 PM Bug #906: clustered mds: lchown not setting uid/gid
- I don't think that I ever did manage to reproduce it.
I haven't thought it through much, but it's also possible th... - 08:49 PM Linux kernel client Bug #1109 (Closed): rbd: btrfs crash
- this was on old code.
- 08:47 PM Bug #1122 (Resolved): kclient: async readahead
- Many people now have noticed that sequential read performance is slower than writes. Is this simply a matter of adju...
- 08:46 PM Bug #1121 (Resolved): rados: rados bench read aborts with an error
- Reported by multiple people now on ceph-devel. Probably easy to fix?
- 08:45 PM Feature #1120 (Resolved): qa: gcov metrics
- generate total coverage statistics for the entire qa suite so we can measure overall coverage and improvements. we c...
- 08:44 PM Feature #1119 (Resolved): qa: gcov/lcov html output
- generate browsable lcov pages for individual tests and/or the whole qa suite
- 07:37 PM Revision 5b7c8ae8 (ceph): osd: protect recovery_wq ops with the recovery lock
- We were calling recovery_item.remove_myself() without holding the
recoveryWQ::lock. Naughty naughty!
Signed-off-by: ... - 07:37 PM Revision b3fb58ea (ceph): crushtool: add -v verbose for --test mode
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 01:27 PM Bug #1118 (Resolved): Crash OSD after upgrdae from 0.28.1 to 0.28.2
- I know, "after" is not always "because". :)
I stop cosd, upgrade (by aptitude from yours repository) and start aga... - 01:10 PM Bug #1116 (Resolved): RecoveryWQ assert failure
- commit:5b7c8ae8bdc26e7593323c76527cb37912b9d833
05/29/2011
- 10:55 PM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
- Hi ,
I'am verifing fsstress test with ceph-client master branch:
commit 98cc99822dac96710a8b64bdc2be4eccffc78956
... - 10:00 PM Revision 57ea5020 (ceph): Add content to obsync package
- 09:42 PM RADOS Bug #1017 (Closed): ceph 0.26 ,mkcephfs --crushmap crush.new ,wait for very long time,mds stat i...
- Looks like you need 'chooseleaf' instead of 'choose' in the crush rules.
- 09:42 PM RADOS Bug #1016 (Closed): ceph 0.26,crushmap change,mount fail.
- Looks like you need 'chooseleaf' instead of 'choose' in the crush rules.
05/28/2011
- 04:14 PM Revision 23242045 (ceph): v0.28.2
05/27/2011
- 09:46 PM Revision 7e1de380 (ceph): hadoop: track Hadoop API changes
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:22 PM Revision 232cd6b3 (ceph): rgw: generate random upload id
- 09:05 PM Revision 4ddf8df8 (ceph): SimpleMessenger: allow multiple calls to shutdown
- Fixes a case where radostool crashed on an error shutdown.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 09:01 PM Revision 8490b784 (ceph): common/Thread.h: const cleanup
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:35 PM Revision a0d521b2 (ceph): rgw: fix signing for some requests
- 07:50 PM Revision 818bfd15 (ceph): rgw: serve multipard init upload
- still needs to generate a random hidden object, and use rados locator
for accessing it. - 05:59 PM Revision 7cfb3b6a (ceph): Merge branch 'wip-obsync'
- 04:01 PM Bug #1116: RecoveryWQ assert failure
- Looks as though this patch has helped.
At least this osd has completd rebalancing.
Great! Thanks! - 12:17 PM Bug #1116: RecoveryWQ assert failure
- Okay, checked this out. It turns out that the only function violating the locking was OSD::do_recovery. Simply adding...
- 09:48 AM Bug #1116 (Resolved): RecoveryWQ assert failure
- From Fyodor:...
- 02:30 PM CephFS Bug #1117 (Resolved): mds: rename rollback broken on slaves during replay
- Best I can tell it's just busted. The rollback object contains all the dentries and inodes, but on a slave it's entir...
- 10:03 AM Bug #1052 (Resolved): obsync: add rados backend tests to test-obsync.py
- Implemented.
- 09:39 AM CephFS Bug #1041: standby-replay fails on multi-mds fsstress journals
- Back from vacation, and I'm trying to remember what's still broken here. Looking through my logs:
1) MDS 1 gets requ... - 09:24 AM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
- Jeff Wu wrote:
> static int ceph_set_page_dirty(struct page *page)
> {
> ...............................
> /* dir... - 08:01 AM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
-
static int ceph_set_page_dirty(struct page *page)
{
...............................
/* dirty the head */
spin... - 08:00 AM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
- Hi ,
I attached some of logs at bug #1096 http://tracker.newdream.net/issues/1096.
:ceph-client-fsstress log 1,2,3.... - 08:17 AM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- > OK, thanks. I'll try out 2.6.39 tomorrow. Will keep you informed.
Now running 2.6.39 everywhere on freshly creat... - 04:37 AM Revision 574b58f3 (ceph): mkcephfs: pass config to osdmaptool
- This lets OSDMap::create_simple() see g_conf.osd_pool_default_size when
creating the initial data, metadata, and rbd ... - 04:31 AM Revision d2ab764b (ceph): drop useless cm.txt
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:20 AM Revision 1292436b (ceph): osdmap: take default pool size from config
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
05/26/2011
- 10:18 PM Revision 9e8484e8 (ceph): rgw: handle POST requests for s3
- 10:07 PM Revision 9b8daa92 (ceph): crushtool: update help
- Signed-off-by: Sage Weil <sage@newdream.net>
- 10:00 PM rgw Feature #767 (In Progress): rgw: incremental/large file uploads
- 09:11 PM Revision 6f704e33 (ceph): obysnc: rgw target: validate all users
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:17 PM Revision 22082c4f (ceph): mon: remove pg_temp mappings when we delete pools
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:17 PM Revision ae5bbc7b (ceph): Merge branch 'wip-obsync'
- 08:15 PM Revision e0cbb131 (ceph): test-obsync: test sync directly from s3->rgw
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:12 PM Revision a93c86e5 (ceph): crushtool: fix --add-item weight being zero when parent bucket(s) created
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:21 PM Revision 56d5d959 (ceph): obsync: fix bucket creation through rgw target
- The rgw: target can now create buckets. Add a test.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 06:04 PM Revision 9cefb56b (ceph): Merge branch 'stable'
- 05:25 PM Revision b2c1bff8 (ceph): test-obsync: test big objects, user-defined xattr
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:22 PM Bug #1098 (In Progress): mds never coming "up:active" awaits in "up:creating"
- shyamali mukherjee wrote:
> I have put OSd logfile and journal to ext3. osd data still comes from "btrfs".
>
> I ... - 11:26 AM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- I have put OSd logfile and journal to ext3. osd data still comes from "btrfs".
I have tried atleast about 50 times... - 10:14 AM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- You switched everything over to ext3?
It doesn't look like a user_xattr issue; the cosd daemon will error out and ... - 10:01 AM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- The cosd has blocked on a btrfs bug; it doesn't have much to do with Ceph.
Eventually your cluster should declare ... - 09:50 AM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- Hi Sage,
I know you have closed the issue. But I could not attach the logfile as it is too huge. I have got few li... - 05:19 PM Revision e9eeb161 (ceph): mkcephfs: set rdir for local mon setup
- Fixes: #1113
Reported-by: Bernard Grymonpon <bernard@openminds.be>
Signed-off-by: Sage Weil <sage@newdream.net> - 04:55 PM Revision 5d51b8fd (ceph): init-ceph: ssh
- Another bell/whistle.
Signed-off-by: Sage Weil <sage@newdream.net> - 04:31 PM rgw Bug #1115: rgw allows users to "give away" s3 objects
- I added a test for this to s3-tests. To run it, use:...
- 03:36 PM rgw Bug #1115 (Resolved): rgw allows users to "give away" s3 objects
- The Rados gateway should not allow the owner of an object to be changed through a PUTACL operation. Amazon doesn't al...
- 03:42 PM Bug #1056 (Won't Fix): obsync: add warning when source owner is not the same as dest owner (after...
- The owner of the object we create is determined by the access key and secret key supplied by the user. It can never b...
- 03:37 PM Bug #1046 (Resolved): rgw: changing object owners
- Filed bug #1115 because RGW's behavior does not match Amazon's.
- 03:33 PM Bug #1046: rgw: changing object owners
- The answer to question #1 is no, Amazon does not allow users to "give away" the ownership of objects.
- 03:29 PM CephFS Bug #1114 (Rejected): NFS export extreme slowdown
- Attached is debug mds 20 output.
Below is ceph -w output for a corresponding period.
Time synchronization is < 0.1s... - 03:28 PM Bug #906: clustered mds: lchown not setting uid/gid
- Greg, what did you do before to reproduce this?
- 02:29 PM Bug #1050 (Won't Fix): obsync: implement --filter to allow certain objects in the source to be sk...
- The original reason we wanted this feature was to skip objects with different owners.
This was handled by the creati... - 02:28 PM Bug #1051 (Resolved): obsync: create a librgw to parse binary ACLs generated by RGW
- 01:22 PM Bug #960 (Resolved): obsync: support rados pool "buckets"
- > - sync directly to/from librados
Implmeneted in the rgw: target.
> - copy amazon acl's into same xattr name tha... - 12:53 PM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- Sage Weil wrote:
> Andre Noll wrote:
> > Hm that does not seem to work. I had to compile a kernel with dynamic debu... - 09:32 AM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- Andre Noll wrote:
> Hm that does not seem to work. I had to compile a kernel with dynamic debug enabled,
> but noth... - 02:30 AM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- Hm that does not seem to work. I had to compile a kernel with dynamic debug enabled,
but nothing makes it to the log... - 11:16 AM CephFS Bug #1108: Large number of files in a directory makes things grind to a halt
- Excellent thanks for the tips. It'll have to wait until Tuesday now for testing but I'll report back then. Going to u...
- 10:10 AM CephFS Bug #1108: Large number of files in a directory makes things grind to a halt
- If that turns out to be too unstable for you and you have gobs of RAM for your MDS, you could also bump up the MDS ca...
- 10:20 AM Bug #1113 (Resolved): rdir is not set correctly for the mons
- Fixed in stable branch. BTW in the future please add a Signed-off-by to your patches... see SubmittingPatches file i...
- 09:45 AM Bug #1113 (Resolved): rdir is not set correctly for the mons
- In mkcephfs, rdir is used to keep the config in for remote hosts, and dir is used for localhost. However, when bootst...
- 09:53 AM Bug #1095: run "rados bench 10 seq -p data" print "error during benchmark: -5"
- Actually, the write benchmark should record how many objects are left and the read benchmark isn't supposed to go pas...
- 09:17 AM Linux kernel client Tasks #1112: check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock) + igrab...
- Hi Jeff-
Are there actual cases of this that you're seeing? I've fixed several of these, but I'm not aware curren... - 12:48 AM Revision 05cfb4d5 (ceph): obysnc: fix content-type on RGWStore
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:36 AM Revision 6cf67a26 (ceph): test-obsync: compare_directory now compares xattrs
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
05/25/2011
- 11:50 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Hi ,i git ceph-client master :
commit 35b0ed997b1a49ff73a6110cbd04681467dbe217
Author: Sage Weil <sage@newdream.n... - 07:40 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- I will build the kernel to verify it.thanks.
- 11:45 PM Linux kernel client Tasks #1112 (Resolved): check all igrab at ceph-client,remove deadlock : spin_lock(&inode->i_lock...
- Hi , at igrab function,it has existed the codes: spin_lock(&inode->i_lock);
if coding this:
spin_lock(&inode->i_... - 10:55 PM Revision 4cae0ea8 (ceph): ceph-pybind-test: test embedded NULLs in data
- Test embedded nulls in rados data. Fix a bug in rados.Object.__str__
Signed-off-by: Colin McCabe <colin.mccabe@dream... - 10:49 PM Revision a2d35295 (ceph): obsync: more fixes for RgwStore
- * Fix content-type handling
* add vvprint and use it in Object::equals.
* support RgwStore::prefix
* more tests
S... - 10:48 PM Revision b76874f6 (ceph): pybind/rados: correctly return data with NULLs
- Correctly handle returning data with embedded NULLs in it.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 10:21 PM Revision 970897ce (ceph): pybind/rados.py: throw NoData on ENODATA
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:56 PM Revision 28c30265 (ceph): mds: fix canceled lock attempt
- If client tries to lock a file, has to wait, and then cancels the attempt,
the client will send an unlock request to ... - 09:34 PM Revision 596a3d6a (ceph): librbd: make image contexts threadsafe
- Use refresh_lock to protect the needs_refresh member, and
ImageContext::lock for the header and snapshot metadata.
S... - 09:22 PM Revision d38001c7 (ceph): pybind/rados.py: rados.Object.key should be string
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:58 PM Revision b2554823 (ceph): obysnc: RgwStore: make sure destination users exist
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:36 PM Revision 5d865fb6 (ceph): obsync: fix DST_OWNER
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:33 PM Revision 73e28f2e (ceph): rgw: return EACCES if acl xattr doesn't exist
- 07:05 PM Revision ea76712a (ceph): obsync: Add boto_retries, remove rgw_store.prefix
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:23 PM Revision e3dd77d8 (ceph): librbd: const cleanup
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 06:06 PM Revision 2aa9151e (ceph): librbd: clean up md_oid use a bit
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 06:01 PM Revision 0adaa6b6 (ceph): rados python bindings: handle xattrs with NULL
- Handle extended attributes that contain NULL bytes correctly, rather
than treating everything as zero-terminated C st... - 05:54 PM Revision d4bfd964 (ceph): PG: fix race in _activate_committed
- Previously, _activate_committed would access the osdmap epoch racing
with handle_osd_map's osdmap update. This would... - 05:50 PM Revision 7de7ba00 (ceph): RgwStore: fix some ACL issues
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:21 PM Revision 42f873e6 (ceph): Proper ACL support for rados targets
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:21 PM Revision 17053739 (ceph): test-obsync: refactor a little bit
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:21 PM Revision e4e098ba (ceph): Rename RadosStore to RgwStore
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:21 PM Revision 3f5f5620 (ceph): test-obysnc.py: support librgw testing
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:38 PM Revision 0aa18f32 (ceph): mds: do not shift to EXCL or MIX while rdlocked
- There was an old change in file_eval() that was allowing us to switch from
SYNC to MIX or EXCL while there were rdloc... - 04:25 PM Messengers Bug #1107 (Resolved): msgr: old outgoing connection + mark_down leaves stale state on remote peer
- 02:57 PM CephFS Bug #1111: file lock requests in wait queue not getting cleaned up after process exit
- pushed kclient fix to ceph-client.git master branch and mds fix ceph.git master branch.
- 02:38 PM CephFS Bug #1111 (In Progress): file lock requests in wait queue not getting cleaned up after process exit
- 12:12 PM CephFS Bug #1111: file lock requests in wait queue not getting cleaned up after process exit
- Adding our test program.
- 11:37 AM CephFS Bug #1111: file lock requests in wait queue not getting cleaned up after process exit
- Here's the notes on our bug and the related PIDs
NODE 1: 192.168.98.112 (client 4113)
- gets the lock and ho... - 11:36 AM CephFS Bug #1111 (Resolved): file lock requests in wait queue not getting cleaned up after process exit
- Our interpretation of events:
1) proc1 requests lock
2) proc1 receives lock
3) proc2 requests lock
4) proc2 gets ... - 12:05 PM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- The kclient isn't responding to a cap revocation message. I'm not seeing anything since 2.6.38 that would have fixed...
- 11:39 AM CephFS Bug #1110: mds: ls -l hangs on concurrent writer
- mds log created with
ceph mds tell 0 injectargs '--debug-mds 20 --debug-ms 1'
The problem occurs at 2011-05... - 10:49 AM CephFS Bug #1110 (Resolved): mds: ls -l hangs on concurrent writer
- ...
- 10:57 AM Bug #1098 (Closed): mds never coming "up:active" awaits in "up:creating"
- 10:56 AM Bug #1012 (Rejected): Autotest: Measure RADOS IO performance under read and write loads
- 10:56 AM Feature #948 (Rejected): autotest: graph rbd performance
- 08:59 AM CephFS Bug #1108: Large number of files in a directory makes things grind to a halt
- Enabling directory fragmention should fix this.. add
mds bal frag = true
to your [mds] section and restart the... - 04:18 AM CephFS Bug #1108 (Closed): Large number of files in a directory makes things grind to a halt
- Whilst extracting a copy of our mail directories onto a 10 node cluster(3xmds, 3xmon, 10xosd) I found that there was ...
- 08:55 AM Linux kernel client Bug #1109 (Closed): rbd: btrfs crash
- ...
- 04:14 AM Revision fe955881 (ceph): crushtool: clean up add-item a bit; don't add item to same bucket twice
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:05 AM Revision dd89ff44 (ceph): crushtool: fix remove-item
- Scan all buckets instead of doing a tree traverse.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 03:30 AM Revision 1c334d1a (ceph): radosgw_admin: update clitest
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 01:58 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Fixed!
Pulled in the latest changes, recompiled, and works like a charm now.
- 01:16 AM Revision ab01d74e (ceph): mkcephfs.in: print out usage if no actions given
- If the user didn't specify any actions, print out a usage message rather
than silently exiting.
Signed-off-by: Colin... - 12:53 AM Revision f7ea7c98 (ceph): rgw: Fix RGWAccess::init_storage_provider
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:06 AM Revision c67dd164 (ceph): mkcephfs: error out on bad usage
- Signed-off-by: Sage Weil <sage@newdream.net>
05/24/2011
- 11:40 PM Revision 5532f897 (ceph): make: fix build for rgw
- 11:33 PM Revision 203a43bf (ceph): rgw_admin: clean warning
- 10:58 PM Messengers Bug #1107 (Resolved): msgr: old outgoing connection + mark_down leaves stale state on remote peer
- Peers A and B:
- A reached epoch 10, in which A and B should exchange heartbeats
- A advances to epoch 20, and ... - 10:30 PM Revision 95c594f6 (ceph): Merge commit 'origin/master' into rgw-multiuser
- 09:29 PM Revision ab278b4b (ceph): rgw_admin: add key create
- 09:19 PM Bug #1095 (Closed): run "rados bench 10 seq -p data" print "error during benchmark: -5"
- Hi Jeff-
I think the problem here is just that the read phase is running out of data to read. Let the write phase... - 09:17 PM Revision bd0eb9a3 (ceph): rgw_admin: subuser and key removal
- 08:38 PM Revision 0566de49 (ceph): Let callers specify that some arguments should not be quoted.
- This lets you do things such as "test -e /foo && bar" or
"cd /tmp && blah". Remember that shell pipelines do not dete... - 08:29 PM Revision dad0a67a (ceph): Simple unit tests for shell quoting.
- 08:27 PM Revision be28e5bf (ceph): Refactor to extract shell quoting into utility function.
- 08:16 PM Revision 1a459dd7 (ceph): Depend on Paramiko 1.7.7 or newer to be able to read modern OpenSSH keys.
- 08:16 PM Revision 7330c3c4 (ceph): journaler: tolerate ENOENT when prezeroing
- ENOENT is okay and expected.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 08:12 PM Revision 37c94af8 (ceph): Pyflakes cleanup.
- 08:07 PM Revision 5d5b1795 (ceph): Add a utility function run.wait to wait for processes to exit.
- 08:06 PM Revision 073a4bbc (ceph): Paramiko ChannelFile.close() didn't actually close the remote stdin.
- Add a wrapper that does the calls shutdown on the channel itself,
to actually cause EOF. Add integration test using r... - 08:01 PM Revision 6dd4774f (ceph): Log debug info of commands actually executed.
- 08:01 PM Revision 9c42fe6b (ceph): Cleanup dead code.
- 08:00 PM Revision f10668f5 (ceph): Allow easy writing to stdin of remote processes.
- 07:36 PM Revision bb13c92a (ceph): test_common.sh: skip rm before put
- The rm before the put is unecessary and actually incorrect now.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.... - 07:34 PM Revision e42736ae (ceph): radostool: rados put should use write_full
- If "rados put" uses write instead of write_full, the resulting object on
the server may be a mismash of old and new o... - 07:22 PM Revision cfe372ec (ceph): Merge branch 'wip_ceph_context'
- 07:21 PM Revision 9ff7cc7c (ceph): Create a libcommon service thread
- Create a libcommon service thread. Use it to handle SIGHUP.
Handle it by means of a flag that gets set. Using a queu... - 05:00 PM Revision 29702685 (ceph): librados: len should be size_t
- Unsigned, and size_t because it's a buffer size.
Fixes signedness warning in testrados.
Signed-off-by: Sage Weil <s... - 04:47 PM Revision ce04e3db (ceph): osd: add ability to explicitly mark unfound as lost
- Instead of automatically marking unfound objects lost (once we've tried
every location we can think of), do it when t... - 04:42 PM Revision 87309e94 (ceph): osd: make automatically marking of unfound as lost optional
- We may not want to do this automatically until we have more confidense in
the recovery code. Even then, possible not... - 04:27 PM Revision cea7b651 (ceph): mds: clean up get_or_create_stray
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:24 PM Revision 081acc4c (ceph): mds: initialize stray_index on startup
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:17 PM Revision 754cfaac (ceph): Merge branch 'stable'
- 01:21 PM CephFS Bug #1104: Segmentation fault when deleting a folder
- I'll try it first thing tomorrow, no more access to the machines now - everything is always updated completely on all...
- 12:55 PM CephFS Bug #1104: Segmentation fault when deleting a folder
- cherry-picked commit:7330c3c473aa128b1e3ecb8752278f655bc79620 to stable. i'm a bit surprised you're seeing this on t...
- 12:50 PM CephFS Bug #1104: Segmentation fault when deleting a folder
- There we go:
[Switching to Thread 0x7ffff5574700 (LWP 27162)]
0x00007ffff67c1165 in raise () from /lib/libc.so.... - 12:18 PM CephFS Bug #1104: Segmentation fault when deleting a folder
- I'll have to rebuild everything, "r" it is optimized out in my build. This will take a little longer...
#6 0x0000... - 11:37 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Can you check with gdb to see what the value of 'r' actually is?
- 11:32 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Tried the stable branch (i'm at ce04e3dbaf2383a521b267585a860f772c4cc786), made debian packages, installed it all, st...
- 11:20 AM CephFS Bug #1104 (Resolved): Segmentation fault when deleting a folder
- Yay! Thanks for your help testing. We'll do 0.28.2 in a few days.
- 11:19 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Compiled from last master sources (sorry, forgot switch to stable branch) not have this trouble. Hooray? Maybe it mak...
- 10:15 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Attached! You may have problems if your libraries don't match mine. There are also the autobuilt debian packages th...
- 09:56 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Sage Weil wrote:
> the 'stable' branch has that fix, or you can apply it manually...
Published in your repository... - 09:28 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- the 'stable' branch has that fix, or you can apply it manually...
- 09:23 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Sage Weil wrote:
> Can you try with this patch applied?
It's 0.28.1 or I should compile master branch?
- 09:01 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Can you try with this patch applied?...
- 01:40 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- I can not attach files to this issue.
http://blog.ufm.su/core.zip - core file
http://blog.ufm.su/mds.zip - log fi... - 12:50 PM Linux kernel client Feature #962: d_prune
- 12:50 PM Linux kernel client Bug #851: make dcache readdir with I_COMPLETE work
- 12:50 PM Linux kernel client Bug #850: make NULL lookup using I_COMPLETE work
- 11:33 AM Bug #1099: osd: handle recovery of lost objects
- For the time being I disabled automatic marking of lost objects. That makes dealing when "recovering" them less of a...
- 11:31 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- This is a kclient bug due to multiple threads entering flush_dirty_caps, which is not reentrant due to commit:e9964c1...
- 09:53 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- The 30k lines still doesn't have the last client_request arrival. I bumped the limit to 50mb. Can you grab a bigger ...
- 04:11 AM Revision d66c6ca1 (ceph): v0.28.1
- 12:28 AM Revision 9a660ac9 (ceph): librads, libceph: store CephContext
- Don't use the global g_ceph_context. Instead, store the CephContext in
the structures provided by the library user.
... - 12:28 AM Revision 13aed89e (ceph): Add CephContext
- A CephContext represents the context held by a single library user.
There can be multiple CephContexts in the same pr... - 12:07 AM Revision 1c7b9821 (ceph): Split common_init_daemonize from common_init_finish
- Split off common_init_daemonize from common_init_finish. cfuse is a
daemon that calls common_init_finish, but handles...
05/23/2011
- 11:52 PM Revision 478c6bbc (ceph): rgw_admin: make interface a bit more explicit
- 10:12 PM Revision c167a28d (ceph): rgw: subuser permissions
- 09:58 PM Revision 6360154d (ceph): mon: verify that crush max does not exceed osd max
- - when injecting a new crushmap
- when adjusting osdmap max_osd
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:45 PM RADOS Bug #1106 (Resolved): crush/osd: inconsistent mapping values
- This was because crush max_devices was osdmap.max_osd - 1. Need to add some loud warnings and checks for this.
- 03:24 PM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- The MDS isn't coming up because teh OSD requests aren't completing because btrfs is wedged. Which kernel are you usin...
- 03:16 PM Feature #1105 (Resolved): have multiple access keys per user in rgw
- Although the radosgw_admin interface needs a bit of polishing, it's implemented as of commit:c167a28d73b665f7239f8fe7...
05/22/2011
- 11:25 PM Revision 5d982803 (ceph): crushtool: add --reweight-item <name> <weight>
- Reweight and individual item via crushtool.
Signed-off-by: Sage Weil <sage@newdream.net>
05/21/2011
- 07:55 PM Revision e9754d88 (ceph): osdmaptool: fail --import-crush if crush max_devices > osdmap max_osd
- Crush will spew non-deterministic badness if it walks off the end of
the osd_weight vector.
Signed-off-by: Sage Weil... - 01:16 AM Revision ba7ef845 (ceph): config: delete after new
- Signed-off-by: Sage Weil <sage@newdream.net>
- 12:57 AM Revision 35ee7e64 (ceph): ceph_crypto: add assert_init
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:57 AM Revision 3a2acefe (ceph): common_init: don't init crypto until after fork
- Get rid of the initialize-then-shutdown-crypto hack. We just initialize
crypto once, after it is safe to do so. There... - 12:10 AM Revision 4cc83a68 (ceph): crush: fix signedness warnings
- Signed-off-by: Sage Weil <sage@newdream.net>
05/20/2011
- 11:46 PM Revision 5baef8f6 (ceph): rgw_admin: able to create multiple keys/subusers
- 11:45 PM Revision cc1737bd (ceph): crushtool: --remove-item name
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:45 PM Revision d287ade5 (ceph): crush: fix tree weight accessor, decompile
- Signed-off-by: Sage Weil <sage@newdream.net>
- 11:45 PM Revision 9a14402a (ceph): crush: fix tree bucket encoding
- I wonder how long this has been broken!
Signed-off-by: Sage Weil <sage@newdream.net> - 10:44 PM Revision 127dcde1 (ceph): crushtool: default to hash 0 (rjenkins1)
- Otherwise we get 255 which is undefined and get bad results!
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 10:15 PM Revision 2cf5048f (ceph): rgw: user info structure supports multiple subusers and keys
- 10:15 PM Revision 27c0bce6 (ceph): mon: fix parsing of 'osd foo N ...' commands with multiple ids
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:15 PM Revision 726aebea (ceph): osd: rework peer map epoch caching
- We try to keep track of which epochs our peers have so that we can be
semi-intelligent about which map incrementals w... - 10:15 PM Revision bc960ac1 (ceph): osd: show last_epoch_clean in PG::Info::History printer
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:15 PM Revision bac1021e (ceph): osd: only forget peer epochs if they are down AND no longer heartbeat p...
- If we forget the peer epoch when we see them go down, we won't share the
map later in update_heartbeat_peers() to tel... - 10:15 PM Revision b5ebe6b5 (ceph): msgr: don't close close_on_empty until outgoing messages are acked
- Otherwise, if we close the socket, we may lose in-flight data.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 10:15 PM Revision a51bf3e9 (ceph): osd: more heartbeat rework
- A few things:
- track Connection* instead of entity_inst_t for hb peers
- we can only send maps over the cluster_me... - 10:15 PM Revision e3191b7d (ceph): osd: merge history when primary sends replica new pg info
- This, among other things, lets us update last_epoch_started and
last_epoch_clean.
Signed-off-by: Sage Weil <sage.wei... - 10:15 PM Revision c22aca1f (ceph): osd: small cleanup
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 10:15 PM Revision 4a83de18 (ceph): osd: update last_epoch_clean in PG::Info::History::merge()
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:27 PM Revision 68021ce8 (ceph): dout: reopen log files on SIGHUP
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:23 PM Revision 277dc66f (ceph): dout: reopen log files on SIGHUP
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 07:39 PM Revision 960d2a36 (ceph): Add SignalSafeQueue
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:29 PM Revision 74691e7c (ceph): osd: clean up old _from target cleanup; fix one case; share map
- Clean up the code to mirror the _to case.
Previously we would not mark down an old _from that is still a _to but wit... - 06:25 PM Revision 0f1be629 (ceph): osd: mark down old _to targets
- If a peer remains a _to target but their address changes, we still want
to mark down the old connection.
Signed-off-... - 06:20 PM Revision 3811d8bf (ceph): osd: share map with old _to peers
- Use new msgr hooks to do this cleanly.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 06:17 PM Revision f87e1dd5 (ceph): osd: clean up handle_osd_ping output
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:12 PM RADOS Bug #1106 (Resolved): crush/osd: inconsistent mapping values
- I'm getting different results for the crush mapping on different nodes. md5sum of the on-disk osdmaps match up. the...
- 05:54 PM Revision 3a7931c7 (ceph): osd: ignore stale requests for heartbeats
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:43 PM Revision f9bea340 (ceph): osd: don't prioritize heartbeat requests
- This could conceivably screw up ordering, and priority doesn't matter
anyway when this is the first message we send t... - 05:42 PM Revision 7a574d88 (ceph): osd: do not clobber explicitly requested heartbeat_to target addresss
- Consider peer P.
- P does down in, say, epoch 60, and back up in epoch 70
- P and requests a heartbeat, as_of 70
- W... - 04:29 PM Revision e1830dbd (ceph): osd: request proper log extent for missing
- We can't blinding ask for everything since last_epoch_started because that
may mean we get some fragment of a backlog... - 03:48 PM Bug #1101 (Resolved): osd: osds don't immediately notice when they've been marked down
- commit:a51bf3e9df027bb9ed58679666ee4207b4185961
- 02:10 AM Bug #1101 (Resolved): osd: osds don't immediately notice when they've been marked down
- I suspect this is related to to the messenger changes (mark_down_on_empty etc). It takes ~20 seconds or more before ...
- 03:44 PM Revision ff031ce8 (ceph): osd: fix log bounds check
- We weren't accounting for the case where we have
(foo,foo]+backlog
i.e., everything is backlog, and rbegin().versi... - 03:35 PM Revision 1dba8dd6 (ceph): osd: osd# is in log entry header/prefix
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:33 PM Revision d75f6237 (ceph): osd: log broken pg state to monitor on startup, activate
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:09 PM Revision b7b8127e (ceph): osd: fix proc_replica_log when peer log is empty
- If the peer log is empty, and we break out of the loop on the first pass,
then clearly last_update has not been adjus... - 02:29 PM Bug #1102 (Resolved): SIGHUP log file reopen is broken
- implemented by commit:277dc66f645f83552789cc6b314f59bdf75ba22d
- 08:04 AM Bug #1102: SIGHUP log file reopen is broken
- on stable branch (v0.28+)
- 08:04 AM Bug #1102 (Resolved): SIGHUP log file reopen is broken
- 02:25 PM Revision f4001108 (ceph): osd: encode keyring as plaintext after --mkkey
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 02:25 PM Revision 93709f89 (ceph): keyring: make encode_plaintext method
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 12:15 PM Bug #1103 (Resolved): osd: log bound mismatch
- commit:e1830dbd09784b7bddf2ab0657b50e6f293cef13
- 08:43 AM Bug #1103 (Resolved): osd: log bound mismatch
- A whole bunch of nodes have inconsistent PG::Info and log bounds. I don't have logs, so I'm not sure how it happened...
- 12:14 PM Bug #1100 (Resolved): osd: marking peers down
- 01:57 AM Bug #1100 (Resolved): osd: marking peers down
- I'm reliably seeing peers mark each other down when they shouldn't on benjamin. There are ~21 osds across 3 nodes, a...
- 10:35 AM Feature #1105 (Resolved): have multiple access keys per user in rgw
- 09:50 AM CephFS Bug #1104: Segmentation fault when deleting a folder
- Logfile from the first mds, as asked:
18:25 < sage> great. add
18:25 < sage> debug mds = 20
18:25 < sage> debu... - 09:47 AM CephFS Bug #1104 (Resolved): Segmentation fault when deleting a folder
- got this after removing a just created folder:
2011-05-20 18:19:09.679553 7f8254c89700 mds0.18 handle_mds_map i am... - 09:05 AM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- Log file is too big.. It can not be attached ( > 5 MB). 192.168.2.101:6800/8459 --> 192.168.2.107 means bz1 ( MDS no...
- 07:45 AM Bug #1099: osd: handle recovery of lost objects
- My hacky workaround was...
- 01:16 AM Bug #1099 (Closed): osd: handle recovery of lost objects
- ...
- 07:41 AM Revision 6995fd51 (ceph): Merge branch 'wip_choose_acting' into stable
- 07:27 AM Revision bdc371e5 (ceph): osd: take remote log when it is clearly superior
- I'm hitting a case where the primary is compensating for a replica's
last_complete < log.tail by sending a log+backlo... - 07:14 AM Revision 4c97cb5f (ceph): osd: fix compensation for bad last_complete
- If the peer has a last_complete below their tail, we can get by with our
log (without backlog) if our tail if _before... - 06:48 AM Revision 332565f1 (ceph): osd: remove some build_prior stringstream cruft
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:46 AM Revision 45e8627c (ceph): osd: remove useless debug print
- We dump this (and more) at the end of the PgPriorSet constructor.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 06:40 AM Revision a2cb690d (ceph): osd: include past acting osds if they were up
- This fixes a bug where we were excluding up (but not acting) nodes from
past intervals, which in turn was triggering ... - 06:38 AM Revision d4b44f9e (ceph): osd: do not exclude me during build_prior
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 06:25 AM Revision f7e6b1c1 (ceph): osd: show final build_prior result
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:46 AM Revision 6f8708ba (ceph): mon: log mkfs as INFO with fs
- The [ERR] log level is misleading.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 01:02 AM Revision dfe52d9e (ceph): OSD, PG: ignore peering messages from before the last peering restart
- Check them before entering the state machine so we can
safely enter the Crashed state on unexpected messages
from the... - 01:02 AM Revision 628665bc (ceph): OSD: decrement message refcount before returning
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 12:46 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- tail -n 30k mds.alpha.log
- 12:20 AM Revision 4404116b (ceph): mds: kick linklock on revoke_stale_caps
- Also use the eval() method and issue caps instead of calling the individual
eval methods.
Signed-off-by: Sage Weil <...
05/19/2011
- 11:15 PM Revision cef8eb9c (ceph): debian: no shlibs:Depends for obsync either
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:15 PM Revision 4e2c1f45 (ceph): debian: no shlibs:Depends for -dev packages
- So says dpkg-gencontrol, at least:
warning: dpkg-gencontrol: Depends field of package librados-dev: unknown substitu... - 11:13 PM Revision 94433898 (ceph): librbd: don't need to link against crypto libs
- All that is done by librados.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 10:59 PM Revision a71981c0 (ceph): PG: add_event, add_next_event: ignore prior_version on backlog events
- We would not have the previous version if we are merging backlog events.
Signed-off-by: Samuel Just <samuel.just@dre... - 10:24 PM Revision 3471d41b (ceph): add ceph_readdir() to libceph
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 10:24 PM Revision 922f7cc3 (ceph): expanding testceph to test open/readdir/telldir
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 10:10 PM Revision f8f6bed6 (ceph): Add run.run option wait, this will make handling stdin easier soon.
- 10:04 PM Revision 3f43c78b (ceph): client: _flush should no-op if nothing to flush
- If there are no FILE_BUFFER cap_refs, then we can bail out early.
Otherwise we will end up dropping refs we don't hav... - 10:04 PM Revision 67533e14 (ceph): client: be more careful with FILE_BUFFER cap refs
- We should either hold a ref or not; whether we release one can't depend on
whether one is held because we can't assum... - 10:04 PM Revision 510f2dd7 (ceph): client: assert(in) on _flush
- We should never arrive in _flush() and not have a reference to the inode
in question, because the presence of dirty b... - 10:04 PM Revision 838067d0 (ceph): client: clean up _flush callers
- Have _flush return true if there are no dirty buffers. Clean up some
redundant conditionals in the callers
Signed-o... - 10:04 PM Revision 3df86c38 (ceph): client: hold FILE_BUFFER ref while waiting for dirty throttle
- We may block in the write path because we've reached out dirty data limit.
Hold a reference to the FILE_BUFFER cap du... - 10:04 PM Revision 8549fc9a (ceph): Merge remote branch 'origin/stable'
- 09:49 PM Revision 8f7d6c7e (ceph): librados: add python bindings for getxattrs
- Add python bindings for getxattrs. Test getxattr, getxattrs, and
setxattr.
Signed-off-by: Colin McCabe <colin.mccabe... - 09:33 PM Revision fe298f64 (ceph): OSD: send a log in response to a log query when the pg dne
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 09:33 PM Revision bc2c31e0 (ceph): PG: choose_log_location: prefer OSDs with a backlog
- Without preferring an OSD with a backlog, PGs would get stuck in the
active state when acting != up and the backlog w... - 08:47 PM Revision 93c2e17c (ceph): Return a structured result from run.run, to make capturing stdout/stder...
- 08:27 PM Revision 9a5c959b (ceph): Add integration tests for signals and connection loss.
- 08:05 PM Revision df84f4e0 (ceph): Check for errors on remote commands.
- 07:46 PM Revision 57f423ba (ceph): librados: add rados_getxattrs API for C bindings
- Support getxattrs in the Rados C API.
Also add a test of getattrs to testrados.c
Signed-off-by: Colin McCabe <colin... - 07:24 PM Revision bcbcf302 (ceph): ReplicatedPG: wait_for_missing_object in _rollback_to
- Previously, we failed if the relevant clone had not yet been recovered.
Signed-off-by: Samuel Just <samuel.just@drea... - 07:17 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Hi ,
The whole thing mds log size is too large, it about 5.4GB , can't attach it to this web.
this web limit the a... - 10:08 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Do you have a larger piece of the mds log you can attach? (Perhaps the whole thing?)
- 07:17 PM Revision 87d7192c (ceph): Add setup.py, install in devel mode into virtualenv.
- 07:16 PM Revision 492fa488 (ceph): Don't close file after copying stdout/stderr to it.
- If a caller uses StringIO to capture the output, they
cannot call .getvalue() after the close.
This also lets you co... - 06:49 PM Revision 40430595 (ceph): testrados: retab with C-style tabs
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:48 PM Revision 6a580bf2 (ceph): testrados: more getxattr / setxattr tests
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:33 PM Revision 1dd17431 (ceph): Remove libcrush from packaging
- This removes libcrush from the packaging system, now that it's been
merged into libcommon.
Signed-off-by: Colin McCa... - 05:20 PM Revision 37df5b1c (ceph): Refactor remote running to support more use cases.
- 05:19 PM Revision 5bfcec26 (ceph): Add debug logging to monkeypatching.
- 05:18 PM Revision 1ed70d78 (ceph): Silence paramiko transport logging.
- 05:17 PM Revision b397eb5d (ceph): Silence a Paramiko crypto deprecation.
- 05:17 PM Revision 85a28a23 (ceph): Make monkeypatching respect order.
- 05:14 PM Revision f16903d7 (ceph): client: do not retake lock in sync_write_commit
- We already hold the lock from a few frames up the stack (ms_dispatch).
Reported-by: Simon Tian <aixt2006@gmail.com>
... - 05:13 PM Revision ce7f78d0 (ceph): ceph.spec.in: fix obsync description
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 04:41 PM Revision 4d39f1be (ceph): journaler: ENOENT is okay on trim
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:29 PM Revision ecb7c961 (ceph): mkcephfs: pick rdir based on whether current daemon is local or not
- We need to pick $rdir as local or remote inside the for name loop.
Fixes: #1094
Signed-off-by: Sage Weil <sage@newdr... - 03:49 PM Bug #1098: mds never coming "up:active" awaits in "up:creating"
- 2011-05-18 16:06:18.643599 41ece940 -- 192.168.2.101:6800/8459 --> 192.168.2.107:6812/18794 -- osd_op(mds0.1:10 604.0...
- 11:30 AM Bug #1098 (Closed): mds never coming "up:active" awaits in "up:creating"
- After upgrading to ceph 0.27 and latest ceph-client-standalone tree I am unable to mount FS. Intial debugging in kern...
- 03:04 PM CephFS Bug #1087 (Resolved): userspace Client readdir_r failing
- 02:45 PM CephFS Bug #1097 (Resolved): client: failed assert in Client::sync_write_commit
- commit:f16903d724150ce7ec6886972a1726509bdcb828 and commit:67533e14439e9b23ee4be5d62277bba6cd99895c
- 09:39 AM Bug #1094 (Resolved): "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.*...
- Thanks for testing!
- 09:35 AM Bug #1094: "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.**** dire...
- Sage,
Thanks! There is one more thing I had to change. But I see that it is fixed in your latest code.
maxosd=`$... - 12:04 AM Revision 5d161aa0 (ceph): PG: make choose_acting a bit smarter
- This change allows old strays that don't need backlogs
to stay acting until current members of the up set are caught ... - 12:04 AM Revision 8c6ce348 (ceph): osd: clean up choose_acting output
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 12:04 AM Revision 9b979797 (ceph): PG: prefer log with longer tail
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 12:04 AM Revision 0aeb8efb (ceph): PG: merge_log- fix extend log case
- Previously, when extending an empty log with a log with the same
last_update, we would fail an assert since we would ... - 12:04 AM Revision dbb2c383 (ceph): PG: _remove_pg, reset info.last_update and info.log_tail on log zero
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 12:04 AM Revision 51daa435 (ceph): PG: choose_acting: we need best_info to have a backlog, not the primary
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 12:04 AM Revision 92706af3 (ceph): PG: reset pg_trim_to in clear_primary_state
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 12:04 AM Revision 524ab3a6 (ceph): PG: GetLog: don't fail if we get an outdated log
- If we request a log from one osd, and then another member of our prior
set comes up with a later last_update, we shou... - 12:04 AM Revision cad3dfae (ceph): PG: choose acting set and newest_update_osd based on a map of all osds
- newest_update osd should be stable when the primary changes, to
prevent cycles of acting set choices. For the same re... - 12:04 AM Revision 2452d415 (ceph): PG: include ourselves in the prior set
- All acting OSDs should be in the prior set, since any of them may have
the newest update.
Signed-off-by: Josh Durgin... - 12:04 AM Revision 2a0f0cd1 (ceph): PG: remove unused argument to adjust_need_up_thru
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
05/18/2011
- 10:37 PM Revision d0752e81 (ceph): Merge branch 'move_crush_to_libcommon'
- 10:37 PM Revision 14a3f262 (ceph): Move crush into libcommon
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:46 PM Revision 0535e4df (ceph): Initial import
- 09:14 PM CephFS Bug #1087: userspace Client readdir_r failing
- Yes.
I have added 'Client::readdir()' and ceph_readdir(), which call Client::readdir_r etc underneath.
This is work... - 04:36 PM CephFS Bug #1087: userspace Client readdir_r failing
- Were you able to sort out the callback return value stuff?
- 08:39 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Hi ,
single mds and two mds , both of them fsstress test hang.
ceph.conf and single mds test log mds.alpha.log atta... - 05:04 PM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Was this with a single mds? Fsstress is known to turn up clustered mds bugs
- 02:51 AM Linux kernel client Bug #1096: LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- do the following steps , be easy to reproduce
$./fsstress -d /mnt/ceph/mdstest -f write=freq -l 100 -n 10000 -p ... - 02:46 AM Linux kernel client Bug #1096 (Resolved): LTP fsstress test always hang ,ceph 0.27.1+linux-2.6.38.6
- Hi ,
I do fsstress test for ceph 0.27.1 + linux-2.6.38.6 + ubuntu 10.10
$modprobe libceph
$modprobe ceph
$mount... - 07:18 PM Revision d4588bae (ceph): Merge branch 'stable'
- 06:48 PM Revision 2fc13de1 (ceph): Move crush into libcommon
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:45 PM Revision 0d79f1de (ceph): man: update cosd man page
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:17 PM Revision 071881d7 (ceph): v0.28
- 05:15 PM Revision b060f5c8 (ceph): Revert "Makefile.am: link some utils with libcrush"
- This reverts commit c26649861e4c154b1bedf6801342d0a8461a2d0a.
I'm not having any problems linking. I suspect this w... - 05:09 PM Revision f1c82aae (ceph): logclient: get rid of send_log; simplify monitor special casing
- Change the SYNC flag to MON and send the Mlog synchronously in the do_log
call. This eliminates teh send_log vestiga... - 05:07 PM Revision baba0a7a (ceph): msgr: fix signedness in alloc_aligned_buffer
- Signed-off-by: Sage Weil <sage@newdream.net>
- 05:01 PM Revision bd1995c2 (ceph): logclient: log synchronously to syslog
- This is simpler. And there is no reason to delay logging to syslog.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:58 PM Revision 4237da88 (ceph): logclient: send entries once per mon session
- We have a lossless session with the monitor! Only send log entries once.
Otherwise, if the mon is down or something,... - 04:27 PM Revision 38ba4762 (ceph): crush: fix clitest now that leading spaces are stripped
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:22 PM Revision 883a980a (ceph): Makefile: don't clean up some files
- 02:36 PM Bug #1094: "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.**** dire...
- I see the problem. Can you please test commit:0efd51dede578e2cc8c68e1a55d1468a06eef83e (the wip-mkcephfsb branch) and...
- 01:43 PM CephFS Bug #1097: client: failed assert in Client::sync_write_commit
- 08:56 AM CephFS Bug #1097 (Resolved): client: failed assert in Client::sync_write_commit
- 2011/5/17 Simon Tian <aixt2006@gmail.com>:
> Hi folks,
>
> ? ? ? When I write and read a file in client A, open wi... - 12:25 PM Revision c2664986 (ceph): Makefile.am: link some utils with libcrush
- 11:53 AM Revision e3841dc6 (ceph): Makefile: don't clean up some files
- 11:25 AM Linux kernel client Bug #1071 (Resolved): rbd: mkfs.ext4 doesn't complete (but mke2fs -j does)
- 10:20 AM Bug #943 (Resolved): 3-mon cluster won't start
- Ok, this should be fixed by commit:4237da886e61c88935d7fb856b49a2d9676cbf9d. Subsequent patches have some further cl...
- 05:04 AM Revision 2b729875 (ceph): Merge remote branch 'origin/stable' into next
- 05:00 AM Revision 2f9ff022 (ceph): page: redefine PAGE_* macros
- Saw this on sid i386:
msg/SimpleMessenger.cc: In function 'void alloc_aligned_buffer(ceph::bufferlist&
, int, int)':... - 05:00 AM Revision 09810cb2 (ceph): page: fix #ifdef guard
- Signed-off-by: Sage Weil <sage@newdream.net>
- 04:29 AM Revision ee7fa813 (ceph): mds: do not shift to EXCL or MIX while rdlocked
- There was an old change in file_eval() that was allowing us to switch from
SYNC to MIX or EXCL while there were rdloc... - 04:08 AM Revision 9be71938 (ceph): vstart: simplify mds keyring add
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:08 AM Revision 812ce6e9 (ceph): Merge branch 'next'
- 01:49 AM Revision 8ad346a3 (ceph): mon: 'auth caps <name> [svc value [svc2 value2 [...]]]'
- Avoid having to futz with cauthtool if possible.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 01:49 AM Revision bfca7ac5 (ceph): osd: add --mkkey mkfs option
- Optionally generate a new key as part of the mkfs step. This makes life
a bit easier for the admin.
Signed-off-by: ... - 12:45 AM Revision 660e6d52 (ceph): Merge remote branch 'origin/next'
- 12:30 AM Revision a22511db (ceph): PG: update same_acting_since when acting or up changes
- This is a hack since we currently use same_up_since to denote the beginning of an interval.
We should probably change... - 12:27 AM Revision 73b99163 (ceph): msgr: avoid clearing connection_state on pipe replacement
- read_message and write_message both dereference connection-state, so avoid
clearing it when replacing a pipe.
read_m... - 12:27 AM Revision 45494b4d (ceph): crushtool: strip leading spaces from identifiers
- No idea where these are coming from! Weird.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 12:27 AM Revision 50be4c46 (ceph): crush: allow - and _ in crushmap type/item names
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 12:05 AM Revision ec63ec3e (ceph): mon: 'osd tree [epoch]'
- Dump crush map + osd state, displayed as a tree.
Signed-off-by: Sage Weil <sage@newdream.net> - 12:05 AM Revision 108b2a69 (ceph): osdmaptool: print crush tree + osd state
- Output osd state combined with crush tree placement. Note osds in tree
that do not exist and list osds that exist th...
05/17/2011
- 11:29 PM Revision 0e3f0923 (ceph): librgw: be quiet by default
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:12 PM Bug #1095 (Closed): run "rados bench 10 seq -p data" print "error during benchmark: -5"
- Hi ,
ceph 0.27 ,
run "rados bench 10 seq -p data",sometimes , print "error during benchmark: -5".
================... - 10:04 PM Revision b5726e11 (ceph): librgw: make API reentrant
- By passing in the configuration, we can use multiple librgw instances in
parallel-- or will be able to, once g_conf i... - 10:04 PM Revision d4c4fe81 (ceph): pybind/rgw: fix python bindings for librgw
- Use string_at to convert between librgw buffers and Python strings.
Signed-off-by: Colin McCabe <colin.mccabe@dreamh... - 09:59 PM CephFS Bug #791 (Resolved): ls -al waits for writes to complete
- commit:ee7fa813ef29890557f0b03bd3950d422484215d
- 04:48 PM CephFS Bug #791: ls -al waits for writes to complete
- I reproduced some long stalls (~20 seconds) due to the loner flip-flopping. Need to analyze the logs (currently on v...
- 09:53 PM Messengers Bug #1093 (Resolved): msgr: race conditon with replaced pipe's connection_state
- commit:73b99163aba7db77aa122eab99780c3d66f0aa91
- 09:03 AM Messengers Bug #1093: msgr: race conditon with replaced pipe's connection_state
- I was unclear: only one of the OSDs died due to this race. Running 10 on one disk just made this kind of race more li...
- 09:26 PM Revision 28e175d6 (ceph): debian: obsync
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:26 PM Revision 34ffe738 (ceph): ceph.spec.in: add obsync
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:26 PM Revision 6d56c20f (ceph): obsync: no .py
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:19 PM Revision bbb1747a (ceph): PG: Replicas send Notifies in response to queries
- Replicas only send Infos during activate.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 09:12 PM Revision 4320cb15 (ceph): Merge branch 'wip-crush'
- 09:09 PM Revision 9c8f30f1 (ceph): PG: choose_log_location, fix error when scanning up set
- ++up.begin() does not skip the primary. Primary might not be up[0].
Signed-off-by: Samuel Just <samuel.just@dreamhos... - 08:58 PM Revision d90458a9 (ceph): osdmap: set type 0 to 'osd'
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:58 PM Revision f6dc19e3 (ceph): crushtool: fix error handling for adding devices
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:35 PM Revision c73e37b0 (ceph): crushtool: fix unittest map
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:33 PM Revision c9a257e3 (ceph): crushtool: fix usage
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:30 PM Revision d40010bd (ceph): crush: add add_item and reweight functions
- Insert a device at a particular point in the hierarchy, and adjust weights
as appropriate.
Signed-off-by: Sage Weil ... - 08:30 PM Revision e46804bb (ceph): osdmap: use straw buckets everywhere by default
- We were using uniform for the leaf buckets. Use straw instead.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:03 PM Revision 042139d1 (ceph): crushtool: include cumulative bucket weight in decompile
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:03 PM Revision 9a2def6e (ceph): crush: fix up constness some
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:04 PM Revision 36fb0846 (ceph): Add Python bindings for librgw
- Add some Python bindings for librgw.
Also add some more verbose error logging to librgw.
Signed-off-by: Colin McCabe... - 04:58 PM Revision b13bbb06 (ceph): PG: PG can receive a log in WaitActingChange requested in GetLog
- Discard logs requested during GetLog once we are in WaitActingChange.
Signed-off-by: Samuel Just <samuel.just@dreamh... - 03:44 PM Bug #943: 3-mon cluster won't start
- This tarball contains the huge logm file I had to drop to recover the cluster in the logm.dropped directory, as well ...
- 03:04 PM Bug #943 (In Progress): 3-mon cluster won't start
- Can you attach a tarball of one of the mon directories with the big files? It's possible this is a side effect of a ...
- 02:27 PM Bug #943: 3-mon cluster won't start
- It happened again, even on the Gbps network. After two mons failed, the third kept accummulating messages in logm fo...
- 02:55 PM Bug #1094: "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.**** dire...
- It is happening due to check_host fails to identify this host as "localhost"
Here is what happened:
On a dif... - 12:36 PM Bug #1094: "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.**** dire...
- 12:35 PM Bug #1094: "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.**** dire...
- can you run mkcephfs with -x (bash -x mkcephfs <regular args>) so we can tell exactly what it's doing?
- 11:24 AM Bug #1094 (Resolved): "mkcephfs -c /etc/ceph.conf --allhosts --mkbtrfs" finds /tmp/mkcephfs.*...
- I have used ceph0.23 for quite sometime. But now after a fresh install and build of ceph 0.27.1
I see that during ... - 02:52 PM Feature #1089 (Resolved): obsync: deb/rpm package
- 01:52 PM RADOS Feature #433 (Resolved): improve osd reweighting
- commit:4320cb15d4840c88b6e5c91c9923fb82749f78f4
- 01:39 PM Revision 8ed372c9 (ceph): rgw: ahrm.. now really fix logging
- 01:18 PM Revision 0b6cb47d (ceph): rgw: fix logging
- 11:24 AM Revision b7b47a02 (ceph): rgw: fix typo
- 11:15 AM Revision 8836b844 (ceph): rgw: don't log operations on unexisting bucket
- 10:57 AM Feature #1091: librados: support pgls filter
- > I assume that we can define a base class for the pgls_filter iterators and
> specialize it according to the filte... - 01:29 AM Feature #1091: librados: support pgls filter
- The original librados list_filter() had the following:
void Rados::list_filter(Rados::ListCtx& ctx, bufferlist& fi... - 09:27 AM Bug #1088 (Closed): osd: assert(is_up) failed when sending queries
- 09:26 AM Bug #1079 (Closed): pgs stuck peering or degraded
- 05:00 AM rgw Feature #1027: rgw log operations on non-existent bucket
- Starting at commit:8836b8447a3a70fc6dd647d070d763f283084ee7 we don't log operations to unexisting bucket. Still need ...
- 12:12 AM Revision e0439626 (ceph): obsync: preserve user-defined metadata
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
05/16/2011
- 11:51 PM Revision 24233f21 (ceph): obsync: filestore: separate xattr metadata nspace
- User-defined s3 metadata lives in a separate namespace from regular S3
metadata like Content-Type, etc.
Signed-off-b... - 11:26 PM Revision 265ab992 (ceph): PG: Don't use exit to call proc_master_log
- exit is also invoked when transitioning to Reset...
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 10:56 PM Revision f863862c (ceph): obysnc: preserve Content-Type
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:47 PM Revision a4bd854f (ceph): client: update ctime for auth, xattr
- This mirrors the kclient fix in d8672d64. The client can have a newer
ctime due to auth or xattr excl caps. This fi... - 09:25 PM Revision 8e6b53fe (ceph): obsync: FileStore: test storing ACLs in xattrs
- Update unit tests now that we're storing ACLs in xattrs. Fix a bug.
Signed-off-by: Colin McCabe <colin.mccabe@dreamh... - 09:12 PM Messengers Bug #1093: msgr: race conditon with replaced pipe's connection_state
- Wow, that's unexpected. If you look at the source you'll notice that the connection_state is referred to in Pipe::wri...
- 05:38 PM Messengers Bug #1093 (Resolved): msgr: race conditon with replaced pipe's connection_state
- When a non-lossy connection is replaced, the messenger sets its connection_state to NULL while holding the pipe_lock....
- 08:42 PM Revision 3865ca56 (ceph): mon: health WARN if monitor quorum is incomplete
- Signed-off-by: Sage Weil <sage@newdream.net>
- 06:02 PM Feature #1091: librados: support pgls filter
- The Objecter is client-side. The filtering is done by the OSD, so as to preserve network bandwidth and reduce client ...
- 04:54 PM Feature #1091: librados: support pgls filter
- Yeah. If the filtering were done on the librados side, there would be little point to the API.
However, even thoug... - 10:30 AM Feature #1091: librados: support pgls filter
- Note that the filtering is being done on the osd side.
- 10:22 AM Feature #1091 (Duplicate): librados: support pgls filter
- pgls_filter support was removed in the librados API redesign while we were converting everything to iterators. I stil...
- 05:54 PM Revision a82e062e (ceph): obsync: FileStore: store ACLs in xattrs
- Store the ACL XML in extended attributes rather than in side files.
Signed-off-by: Colin McCabe <colin.mccabe@dreamh... - 05:42 PM Revision ac6afe06 (ceph): obsync: FileStore: test for xattr support
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:17 PM Bug #1051: obsync: create a librgw to parse binary ACLs generated by RGW
- librgw exists and is passing unit tests. The thing to do now is to get the python bindings to librgw working, and the...
- 05:16 PM Bug #1040 (Resolved): obsync: preserve content-type, misc metadata
- implemented preserving user-defined metadata by commit:24233f210e1391454ae02140b65da41a30f71209 and commit:e043962601...
- 04:00 PM Bug #1040: obsync: preserve content-type, misc metadata
- We now preserve content-type by commit:f863862c67bd043ba8a7e61b4d776a0bd7ae924c
I'm working on preserving the othe... - 02:48 PM Bug #906: clustered mds: lchown not setting uid/gid
- audit of the uclinet vs kclient code turned up one difference, but it was a bug fix in kclient that was missing from ...
- 01:24 PM Revision 1db29a26 (ceph): rados: don't force order on params
- 11:12 AM Feature #1092 (Rejected): mon: checkpointing
- ability to checkpoint monitor state to facilitate rollback. To be used in combination with #1080.
- 11:11 AM Feature #1080 (Resolved): osd: cluster snapshot
- going to call mon checkpointing out of scope for now. we can that later as needed.
- 11:10 AM Bug #1085 (Won't Fix): bug in cclass
- cclass will be gone in v0.28, which will be out in the next day or two!
- 06:26 AM Bug #1090 (Resolved): broken param parsing in the rados tool
- Fixed, commit:1db29a261016e64f2fba65d3b911991fa29f3d40.
- 06:14 AM Bug #1090 (Resolved): broken param parsing in the rados tool
- 'rados ls -p data' does not return what 'rados -p data ls' returns.
- 03:56 AM Revision e93c0fc0 (ceph): fix segfault introduced by commit de640d85fa3e0e5e5a31704eab5a8714a1ffe867
- That commit introduces the line 'cur_con->put()' which has the possibility
of being called while cur_con is not initi...
05/15/2011
05/14/2011
- 09:07 PM Revision cd75a9d2 (ceph): osd: lazily close connections to down peers
- If we hear from a peer that should be dead, tell them, but mark our
connection so that it will close after that messa... - 09:07 PM Revision a5b5aea4 (ceph): msgr: mark_down_on_empty and mark_disposable
- Mark a connection to close when messages are sent, and to close on any
error. We can use this to tell people who sho... - 08:46 PM Revision 5ecc42b5 (ceph): PG: Remove downed osds from peer_missing and peer_info
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 08:42 PM Revision ba753630 (ceph): PG: Only pull the master log from a member of the prior_set
- There must be a member of the prior_set such that no other
osd has a more recent last_update. This way, prior_set_af... - 12:26 AM Revision 6af0379e (ceph): rgw: Move rgw_log_level to md_config_t
- Need to do this to get librgw to be usable as a standalone library
without unresolved symbols. Also, this makes it co... - 12:16 AM Revision 56cab8ca (ceph): Makefile.am: add SimpleMessenger.cc to libcommon
- libcommon depends on this file, and there's no other library that it
could go in. It is certainly silly to manually i... - 12:10 AM Revision 924c000b (ceph): librgw: only include rgw_acl.cc and librgw.cc
- Rather than putting all of RGW into librgw, only put rgw_acl.cc. Have
RGW use librgw instead of re-including the same...
05/13/2011
- 11:39 PM Revision 298e5c72 (ceph): rgw_acl: move constructors, destructors to .cc
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:28 PM Revision a17db024 (ceph): MDS: don't journal slave ops if we only have caps.
- Previously we wanted to journal if we had caps on something. Now
that we're being strict about only journaling stuff ... - 11:28 PM Revision b8ddecce (ceph): MDS: do journal on rename if we're auth for the inode.
- We missed this case: we can be auth for the inode being moved without
being auth for the srcdn (first case) or owning... - 11:28 PM Revision e8504c0b (ceph): uclient: do not accept max_size changes unless they're from auth mds.
- Unlike most of the cap options, max_size is an inode member. This meant
that if we got a shared cap grant from a repl... - 11:08 PM Revision 9847eb8b (ceph): rgw: put XML-to-bin translation into a librgw
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:08 PM Revision 883d1807 (ceph): librgw: small error handling fix
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:08 PM Revision 1c98da66 (ceph): librgw: use dout for logging
- Use dout for logging so that the librgw library user can turn off or
redirect the logs if necessary.
Signed-off-by: ... - 11:08 PM Revision 0dea92f6 (ceph): boto_tool.py: use s3-tests config file
- boto_tool now grabs the configuration variables it needs from the
s3-tests config file, similar to s3-tests and test-... - 11:08 PM Revision 50e41fbe (ceph): boto_tool.py: fix old-style argument-passing
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:08 PM Revision 6f4f702b (ceph): boto_tool.py: add --rmobjects, --rm_rf
- Add some options to help destroy buckets.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 11:08 PM Revision 12deaaa4 (ceph): obsync: add DST_CONSISTENCY
- The DST_CONSISTENCY variable allows us to specify that the destination
is expected to use read-after-write consistenc... - 11:08 PM Revision bf81df27 (ceph): obsync: fix eventual consistency handler
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:19 PM Revision 3d4971b3 (ceph): Merge remote branch 'origin/osd_snap' into stable
- 09:07 PM Revision 6e0e5532 (ceph): PG: search_for_missing takes the other osd's missing set
- Previously, search_for_missing was erroneously passed the
primary's missing in a few places.
Signed-off-by: Samuel J... - 08:56 PM Revision e0d83fe7 (ceph): PG: search_for_missing takes the other osd's missing set
- Previously, search_for_missing was erroneously passed the
primary's missing in a few places.
Signed-off-by: Samuel J... - 08:13 PM Revision 89a821c6 (ceph): radosgw_admin: fix clitest
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:59 PM Revision 8161122b (ceph): fix null deref when callback invoked en route from readdir_r rather tha...
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 04:59 PM Revision 72ca96e1 (ceph): add basic test case for readdir_r
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 04:31 PM Bug #1088 (Closed): osd: assert(is_up) failed when sending queries
- This happened when I was stress testing the peering code with 10 osds running off one disk, streaming writes, and mar...
- 01:51 PM Linux kernel client Bug #1086 (Resolved): rbd: iozone failure
- 01:51 PM Linux kernel client Bug #1086: rbd: iozone failure
- fixed by commit:1fec70932d867416ffe620dd17005f168cc84eb5
- 09:10 AM Linux kernel client Bug #1086: rbd: iozone failure
- i just noticed this comment didn't post yesterday, no wonder yehuda didn't know what i was talking about :)
here:
... - 06:29 AM Linux kernel client Bug #1086: rbd: iozone failure
- The problem is that our use of blk_end_request is wrong, as it assumes ordering on the requests completion. In most r...
- 10:23 AM CephFS Bug #1087: userspace Client readdir_r failing
- Oh, I got it.
I don't really remember how the readdir works at this point, but if you follow the calls for libceph's... - 10:16 AM CephFS Bug #1087: userspace Client readdir_r failing
- Just thought about this some more... what I need to reconcile is any differences between ceph_ll_add_dirent and _read...
- 10:03 AM CephFS Bug #1087: userspace Client readdir_r failing
- Yeah.. sorry about the context... this is in libceph testing.
ceph_readdir_r was already implemented when I looked a... - 09:33 AM CephFS Bug #1087: userspace Client readdir_r failing
- Well, looking at this real quick I see:
29) lookup on readdir_r_test (succeeds, 0)
30) lookup on readdir_r_test/opene... - 07:24 AM CephFS Bug #1087 (Resolved): userspace Client readdir_r failing
- I chased this down a bit of a ways but there's a lot to look through.
This log is output from testceph with client d...
05/12/2011
- 10:43 PM Revision 84644dc5 (ceph): uclient: compare _revoked_ caps when deciding whether to release.
- cap->issued is already set to new_caps, so that branch was never taken!
Signed-off-by: Greg Farnum <gregory.farnum@d... - 10:36 PM Revision 932f4eb0 (ceph): uclient: clear out cap->wanted when caps get revoked.
- This ensures that we will send a response to the MDS letting it know
that we've revoked our caps.
Signed-off-by: Gre... - 09:34 PM Revision 5e2b57d0 (ceph): uclient: be more careful about sending caps.
- This should prevent us from "losing" caps off the dirty list. See
#1063. If we have dirty caps we don't want to short... - 06:01 PM Revision 91a268ed (ceph): radosgw_admin: dump log by object
- instead of only by date+bucket.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:13 PM CephFS Bug #1063 (Can't reproduce): dbench breaks if MDS and client times aren't synced
- I won't be surprised if this comes back again, but I can't reproduce it and there've been several fixes for client ca...
- 03:50 PM Revision 30491e8f (ceph): updated test to cover "." directory stat
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 03:50 PM Revision 4456b6c3 (ceph): Add analogous special case for "." directory alongside ".." in _lookup
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net> - 11:12 AM Linux kernel client Bug #1086: rbd: iozone failure
- My comment.
I have the starting file size 4G (-n4g) because on this server 4G memory. Into smaller files on the se... - 10:06 AM Linux kernel client Bug #1086: rbd: iozone failure
- strangely, the file looks correct (before and after a remount):...
- 10:00 AM Linux kernel client Bug #1086 (Resolved): rbd: iozone failure
- I was able to reproduce Fyodor's problem on rbd (latest kernel) and ext2:...
- 09:43 AM rgw Bug #1083 (Won't Fix): rgw: log by user, user+bucket
- nevermind. we can just list the log objects directly from the .log pool.
- 09:11 AM Linux kernel client Bug #557 (Can't reproduce): BUG_ON(!session->s_num_cap_releases);
- 09:10 AM Linux kernel client Bug #465 (Resolved): need to refresh osdmap when full flag is set
- added bit to subscribe to next osdmap if current osdmap has full bit set.
- 08:55 AM Linux kernel client Bug #1071: rbd: mkfs.ext4 doesn't complete (but mke2fs -j does)
- Blarg, I can't reproduce this consistently. That bisect is probably bogus.
- 08:54 AM Linux kernel client Bug #909 (Can't reproduce): ceph-client+ceph v0.25.1,iozone test, "libceph: tid 115358 timed out...
- 04:21 AM Revision 935f7dc1 (ceph): mds: drop unneed default arg
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:39 AM Bug #1085 (Won't Fix): bug in cclass
- ...
- 12:57 AM Revision 2655a2b5 (ceph): Merge branch 'osd_pgls'
- 12:56 AM Revision a6417c6a (ceph): objecter: set pgls start_epoch field
- For each pg, start out with start_epoch = 0 in the first request. For
subsequent requests, set it to the first reply... - 12:55 AM Revision 8a1644ef (ceph): osd: add pgls start_epoch field
- If the pgls.start_epoch is set, the cookie is only considered valid if the
osd pg interval has not changed since then... - 12:51 AM Revision 222126e8 (ceph): rgw: in S3 PUT, don't crash on Content-Length == 0
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:47 AM Revision 5c382d35 (ceph): objecter: fix calc_op_budget bit mask checks
- Use the helpers; we need to mask out several bits and compare.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 12:45 AM Revision b6cccc74 (ceph): Objecter: switch handle_osd_map op resending around
- We need to order the resend by tid. We could do that in a
set with a special-purpose comparison function, but just
sw... - 12:27 AM Revision 1d29cc7c (ceph): rgw: in S3 PUT, don't crash on Content-Length == 0
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
05/11/2011
- 11:52 PM Revision 5e4f6bae (ceph): Objecter: implement operator<.
- This will maintain ordering of Ops when they're in eg STL sets.
Previously Objecter::handle_osd_map would indiscrimin... - 11:27 PM Revision c155a2b7 (ceph): osd: prepend missing objects to pgls results
- This will prepend any missing objects to the set of objects returned by
a sequence of PGLS operations. Because recov... - 09:52 PM CephFS Bug #1084: blogbench won't finish: waiting for Fr cap forever
- i initially thought something like this would work
diff --git a/src/mds/Locker.cc b/src/mds/Locker.cc
index 3c7... - 08:03 PM CephFS Bug #1084 (Resolved): blogbench won't finish: waiting for Fr cap forever
- Run blogbench with kclient: blogbench -d /mnt/ceph/henry/b5/
Blogbench won't finish and keeps waiting for Fr caps of... - 09:00 PM Revision d9896b3c (ceph): obsync: handle eventual consistency issues
- Handle eventual consistency issues so that obsync will be usable on more
S3 stores.
Signed-off-by: Colin McCabe <col... - 08:58 PM Revision 7083777c (ceph): osd: remove weird commit_op_seq fast-forward
- This doesn't serve any purpose that we can discern.
In fact, it might cause problems because it'd allow the journal ... - 08:58 PM Revision 82f9a923 (ceph): osd: key Missing::rmissing on version (not eversion)
- This switches the key to the uint64_t (version_t) only, which is still
unique given a particular timeline (which is a... - 08:54 PM Revision f1af92fb (ceph): PG: choose_log_location, fix error when scanning up set
- ++up.begin() does not skip the primary. Primary might not be up[0].
Signed-off-by: Samuel Just <samuel.just@dreamhos... - 08:39 PM Revision 326d01b2 (ceph): osd: support rollback to cluster snapshot
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:27 PM Revision 2dc891f6 (ceph): Clock: add new clock_offset config option, and use it in g_clock.now()
- This way we can test clock drifts without needing to actually
drift the clocks.
Signed-off-by: Greg Farnum <gregory.... - 08:11 PM Revision 7aed34c2 (ceph): clock: remove cruft.
- There were some odd pieces that are artifacts of a very old and
different use. Remove them to simplify the interface ... - 08:11 PM Revision 88641b88 (ceph): osd: trigger a store snapshot when the osdmap says to
- Move the OSDMap decoding up a bit so that we can either snapshot or flush.
We can't do it after we take map_lock or e... - 08:10 PM Revision 6db09bac (ceph): filestore: add a snapshot command to create a snapshot of the entire store
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:10 PM Revision 918eeaf0 (ceph): mon: add 'osd cluster_snap foo' command
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:10 PM Revision 2ced4e24 (ceph): osdmap: add cluster_snapshot field
- Add a cluster_snapshot marker in the map that is valid for a single epoch
to do a coordinated snapshot of the entire ... - 07:12 PM rgw Bug #1081 (Resolved): rgw: 500 error with x-amz-acl instead of explicit xml
- yay fixed! patch is in next branch (for v0.28).
- 05:59 PM rgw Bug #1081: rgw: 500 error with x-amz-acl instead of explicit xml
- should be fixed by 1d29cc7c7627683ba0ae2aa064abab4ea942b4e8.
Just need to test. - 04:37 PM rgw Bug #1081: rgw: 500 error with x-amz-acl instead of explicit xml
- Setting canned ACLs works for me in the tests I am running.
I am running more s3-tests, so maybe that will unearth... - 03:43 PM rgw Bug #1081: rgw: 500 error with x-amz-acl instead of explicit xml
- Here's a tcpdump snippet from s3-tests that works, compare against this to find the cause. (But 500 => there's an rgw...
- 03:41 PM rgw Bug #1081: rgw: 500 error with x-amz-acl instead of explicit xml
- When running $Conn->put_bucket_acl('berlertestobsync1', '', { 'x-amz-acl' => 'public-read' }); where $Conn is an S3::...
- 03:41 PM rgw Bug #1081: rgw: 500 error with x-amz-acl instead of explicit xml
- here's the full Response object (it contains the request object, so you can see what produced it)
$VAR1 = bless( {... - 03:22 PM rgw Bug #1081 (Resolved): rgw: 500 error with x-amz-acl instead of explicit xml
- if something like
'_headers' => bless( {
'user-agent... - 06:56 PM Revision d3aa0c1e (ceph): PG: Replicas send Notifies in response to queries
- Replicas only send Infos during activate.
Signed-off-by: Samuel Just <samuel.just@dreamhost.com> - 06:42 PM Revision 151bf29d (ceph): Merge branch 'next'
- 06:37 PM Revision 8d201d4b (ceph): librbd: tolerate ENOENT when trying to delete an object.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 06:37 PM Revision 484e6e6f (ceph): rados_sync: tolerate ENOENT when deleting an object.
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 06:37 PM Revision 140886c8 (ceph): mdcache: check return values when purging an inode.
- Previously we weren't looking, and if there's a problem
we probably shouldn't be moving on.
Signed-off-by: Greg Farnu... - 06:07 PM rgw Bug #1083: rgw: log by user, user+bucket
- actually, we can scratch problem 2.. each log entry apparently has the bucket owner at the time of the request.
- 04:19 PM rgw Bug #1083 (Won't Fix): rgw: log by user, user+bucket
- problem 1:
- user creates bucket at beginning if day
- pumps full of data
- removes bucket at end of day
- we... - 06:06 PM Bug #1030 (Resolved): osd: list pool/bucket contents excludes missing objects
- 06:04 PM Revision 1429d776 (ceph): test-obsync.sh: fix obsync unit tests
- Fix the obsync unit tests to take into account the new ACL changes.
ACLs must be either translated or ignored when co... - 05:28 PM Revision 0f42099a (ceph): expand testceph to check xattrs
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
- 05:28 PM Revision b0e0c361 (ceph): client: support security. namespace
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
- 05:28 PM Revision 3521771c (ceph): support for xattrs in libceph
- Signed-off-by: Brian Chrisman <brchrisman@gmail.com>
- 03:47 PM Feature #1082 (Rejected): obsync: swift support
- 01:48 PM Bug #1079: pgs stuck peering or degraded
- The ones stuck in degraded were likely caused by the bug fixed in f1af92fb3d3bdab5a74ef40744028001d1943203.
- 01:33 PM Feature #1080 (Resolved): osd: cluster snapshot
- create a snapshot of all osds so we can rollback the state of the entire osd cluster
warping back the monitor will... - 01:24 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Scratch that, I did manage to reproduce locally. It just took a bit longer.
- 01:07 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- On the other hand, adding a clock skew option and setting the MDS into the future doesn't let me reproduce the broken...
- 10:10 AM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Well, job 576 completed successfully after TV time-synced the cluster. Looks like bad mtimes are somehow causing the ...
- 11:01 AM Bug #1078 (Resolved): rados remove fails silently on non-existent objects
- Okay, I checked through these. A lot of callers don't pay any attention to the return code from remove but I looked a...
- 10:13 AM CephFS Bug #930 (Resolved): libceph not exporting getattr
- commit:3521771cb6bdb8eb0cbec7dc27a9999ddb494ad0
- 04:35 AM Revision 3a8f36f5 (ceph): journaler: tolerate ENOENT when prezeroing
- ENOENT is okay and expected.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:32 AM Revision d2243e82 (ceph): osd: unlink of nonexistent object should return -ENOENT
- fixes bug #1078.
- 04:31 AM Revision f114cf18 (ceph): monclient: fix crash on shutdown
- cur_con may be null on shutdown.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 12:46 AM Revision 6f340456 (ceph): Revert "osd: unlink of nonexistent object should return -ENOENT"
- This reverts commit a7f87965f5c49fa878dadd458f87b7974252ab6e.
This commit breaks at least how Filer does zeroing. We...
05/10/2011
- 11:58 PM Revision 59995908 (ceph): Merge branch 'wip-merge-radostool-with-radossync'
- 11:54 PM Revision de640d85 (ceph): monclient: maintain explicit session connection; ignore stray messages
- Maintain an explicit Connection handle to send messages and mark_down old
monitor connections. Ignore any incoming m... - 11:45 PM Revision 3425a8e5 (ceph): rados tool: integrate rados_sync with rados tool
- * integrate rados_sync with rados_tool
* Improve rados tool usage a bit
* Rename test_rados_sync.sh to test_rados_too... - 11:25 PM Revision 15756550 (ceph): rados tool: change initial argument parsing a bit
- Use the ceph_argparse functions. Prepare to integrate with rados_sync.
Signed-off-by: Colin McCabe <colin.mccabe@dre... - 10:23 PM Revision 203edaca (ceph): librados: don't crash if we call connect twice
- Fixes: #1034
Reported-by: Wido den Hollander <wido@widodh.nl>
Signed-off-by: Sage Weil <sage@newdream.net> - 09:49 PM Bug #1078: rados remove fails silently on non-existent objects
- probably needs
diff --git a/src/osdc/Journaler.cc b/src/osdc/Journaler.cc
index 9322ad2..6fbe3f8 100644
--- a/sr... - 05:11 PM Bug #1078 (In Progress): rados remove fails silently on non-existent objects
- Yeah, it definitely hits other things. The MDS no longer starts up due to error codes coming back on prezeroing.
I... - 02:23 PM Bug #1078 (Resolved): rados remove fails silently on non-existent objects
- Fixed by commit:a7f87965f5c49fa878dadd458f87b7974252ab6e.
- 01:11 PM Bug #1078: rados remove fails silently on non-existent objects
- This will fix it, however, I'm not sure either whether it was done on purpose or was just an oversight and/or how it'...
- 11:56 AM Bug #1078 (Resolved): rados remove fails silently on non-existent objects
- Trying to remove a non-existent object produces no error. If this is intentional, it needs to be documented in the li...
- 09:33 PM Revision d67dba76 (ceph): Merge remote branch 'origin/stable'
- 09:20 PM Revision a7f87965 (ceph): osd: unlink of nonexistent object should return -ENOENT
- fixes bug #1078.
- 06:16 PM Revision 331c01e8 (ceph): rados_sync: implement --delete-after, fix bugs
- Implement --delete-after for both export and import.
Fix DIR* leaks.
Signed-off-by: Colin McCabe <colin.mccabe@drea... - 06:16 PM Revision 6e55b237 (ceph): rados_sync: support --force
- Support --force, which re-copies all objects all the time.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 06:16 PM Revision 001c18c1 (ceph): test_rados_sync: test --force
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:16 PM Revision 357910c5 (ceph): Allow dashes in ceph_argparse, etc.
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:40 PM Bug #1079: pgs stuck peering or degraded
- It looks like the degraded ones are staying that way because they need backlogs, but we didn't populate peer_backlog_...
- 04:49 PM Bug #1079 (Closed): pgs stuck peering or degraded
- Using the same setup as in #1073, but with 10 osds, the cluster recovered most pgs, but a few were stuck degraded, an...
- 04:50 PM Bug #1077 (Resolved): integrate rados_sync with rados tool
- implemented in 3425a8e5031a4f0c9c0eb85e8a329b02d05b9420
- 11:43 AM Bug #1077 (Resolved): integrate rados_sync with rados tool
- As we discussed.
- 04:36 PM Bug #1033 (Resolved): osd: CephxClientHandler::handle_response
- commit:de640d85fa3e0e5e5a31704eab5a8714a1ffe867
- 03:40 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Okay, hopefully we can rerun this with a time-synced cluster soon and see if that's what is causing the breakage.
... - 03:22 PM Revision d006c6f2 (ceph): osd: initialize oi.oloc if on-disk value is bogus
- If the on-disk locator is undefined (upgrade of an old cluster?) initialize
the oloc fields based on the PG::Info.
R... - 03:04 PM Bug #1034 (Resolved): librados: Calling connect twice causes a segfault
- 12:15 PM Revision f00edf73 (ceph): rgw: switch bucket creation operations
- First we create the pool, then we create the bucket object. This
should have the effect of making the bucket creation... - 11:19 AM Bug #1074 (Resolved): rados_sync: implement --delete
- implemented by 331c01e847c471980c31980a618c3bce3face50e
- 10:47 AM Feature #1068 (Resolved): rados: incremental import/export
- 10:47 AM Feature #1069 (Resolved): rados: support attrs in import/export
- 10:04 AM Bug #1076 (Resolved): avoid sparse read for small reads
- It doesn't make sense to make sparse read for reads that are smaller than the block size. This should either be in li...
- 09:43 AM Bug #1021 (Can't reproduce): osd: _process_pg_info FAILED assert(pg->log.tail <= pg->info.last_co...
- Let's see if this comes up post-refactor.
- 09:43 AM Bug #1028 (Resolved): segfault in OSDMap::object_locator_to_pg
- 04:53 AM Bug #1028: segfault in OSDMap::object_locator_to_pg
- ok, it seems fixed. Now back to #1022
- 12:33 AM Bug #1028: segfault in OSDMap::object_locator_to_pg
- Thank you for the patch, compiling right now.
This is indeed an old FS that got created approximately a year ago, ... - 06:01 AM rgw Bug #1059 (Resolved): RGW consistency issues
- Fixed now, commit:f00edf73284fc0f6e32973d16f58eb81f7b96bf8. However, this might have impact on performance.
05/09/2011
- 11:01 PM Revision 0ac419e0 (ceph): osd: drop bad warning
- The stats won't match reality if there are any missing or if there are any
snapped objects.
Signed-off-by: Sage Weil... - 11:01 PM Revision 8e1e45c0 (ceph): osd: reset last_complete on mark_all_unfound_as_lost if no more missing
- If we marked _all_ missing as lost, reset last_complete, since missing is
now empty!
Signed-off-by: Sage Weil <sage.... - 11:01 PM Revision 70d8c994 (ceph): osd: simplify build_might_have_unfound
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:01 PM Revision 8a781f11 (ceph): osd: fix osd$foo typos
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:01 PM Revision 7a6b9b97 (ceph): osd: fix pollution of peer_info
- The ++ postfix has no effect here! We really want +1.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 11:01 PM Revision 7ada5cd6 (ceph): osd: wait for up_thru updates
- Before the primary can go active we need to wait for the up_thru in the
osdmap to reflect that we were alive during t... - 11:01 PM Revision 6d70592d (ceph): osd: log debug output for Crashed state
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:01 PM Revision 8cb861c3 (ceph): osd: rename Pending -> WaitActingChange
- We only use the Pending state while waiting for the acting set to change.
Rename the state and log it appropriately s... - 10:18 PM Revision d9ea95f2 (ceph): rados tool: remove import/export
- rados_sync replaces rados import / rados export
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 04:46 PM RADOS Feature #1075 (Resolved): crushtool: warn if uniform item weights vary
- uniform bucket weights are ignored (and presumed uniform, hence "uniform").
- 04:29 PM Bug #1073 (Resolved): osd: failed assert: info.last_complete == info.last_update
- fixed by commit:7ada5cd685fcf3cae4e1c5d2dd81ea1817cceee7
- 09:51 AM Bug #998 (In Progress): qemu/librbd race conditon
- Christian Brunner had a similar error, but in aio_write during a yum upgrade. We should do more testing on this.
- 09:05 AM Bug #1028: segfault in OSDMap::object_locator_to_pg
- This problem is that the locator stored in the object_info_t on disk is wrong. Can you say anything about when the o...
- 05:28 AM rgw Bug #1035 (Resolved): incorrect rgw log data
- This is already fixed, commit:a09eb0c33f6b05714bd4f780f79c70cb4529f840.
05/08/2011
- 11:30 PM Bug #1028: segfault in OSDMap::object_locator_to_pg
- Cherry-picked 85292b367b0e6e6d8963de32ad198482500c887f into the stable branch, here are the logs... I kept the core f...
05/07/2011
- 07:56 PM Revision 1cb611a0 (ceph): .gitignore: rados_sync
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 02:15 PM Bug #1073: osd: failed assert: info.last_complete == info.last_update
- we only wait for up_thru updates if some_down:
if (prior.some_down) {
need_up_thru = true;
for (vect... - 01:51 PM Bug #1073: osd: failed assert: info.last_complete == info.last_update
- this is wrong:
2011-05-06 17:33:48.284200 7f1466b56710 osd4 31 pg[0.12( v 21'17 (21'15,21'17] n=17 ec=2 les=11 31/... - 01:39 PM Bug #1073: osd: failed assert: info.last_complete == info.last_update
- something is definitely going wrong here.. i see
2011-05-06 17:34:24.391722 7fe3aae59710 osd9 43 pg[0.12( v 21'17... - 05:10 AM Revision ea0a1395 (ceph): osd: fix compilation for some g++ versions
- wasn't compiling on lenny, g++ ver 4.3.2. Might be that
it's also due to differebt boost version. - 04:59 AM Revision 25bfb987 (ceph): osd: reassert our assert definition after including boostchart
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:59 AM Revision f9ed9885 (ceph): assert: make our assert clobber any others too
- Two can play this game, /usr/include/assert.h!
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 04:42 AM Revision 7db49499 (ceph): rgw: opening bucket io ctx stats bucket info if failed
- this should trigger request of a new osdmap if we were racing
with bucket creation. - 04:42 AM Revision 588fe672 (ceph): rgw: minor cleanup
- 12:43 AM Revision 290668c0 (ceph): Merge branch 'wip-rados-sync'
- 12:42 AM Revision fbe0bd1b (ceph): test_rados_sync: check that second sync does nada
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 12:33 AM Revision 210c38d2 (ceph): rados_sync: more fixes
- * separate BackedUpObject::from_path and BackedUpObject::from_file.
* librados functions return negative values on e...
05/06/2011
- 11:50 PM Revision 85292b36 (ceph): osd: warn when obc locator doesn't match ops
- This will hopefully help us track down #1028.
Signed-off-by: Sage Weil <sage@newdream.net> - 11:41 PM Revision 5a726efc (ceph): rados_sync: fix memory leak, other misc fixes
- * Store BackedUpObject in auto_ptr to avoid memory leaks
* better comment for BackedUpObject::get_fs_path()
* xattr... - 11:26 PM Revision 5f7df9af (ceph): osdmap: refactor mapping; don't assert on missing pool
- We were asserting on a pg with a missing pool fed to pg_to_acting_osds
because raw_to_temp_osds called raw_pg_to_pg. ... - 11:13 PM Revision a61484e6 (ceph): test_rados_sync: test xattr preservation
- test xattr preservation across import/export
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 11:03 PM Revision 6b870bbd (ceph): test_rados_sync: add import/export test
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 11:01 PM Revision b169379c (ceph): rados_sync: complain if import src is inaccessible
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:55 PM Revision 010ffced (ceph): rados_sync: fix --create for import
- Fix import's pool creation faculty.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 10:55 PM Revision e3212711 (ceph): add test_rados_sync.sh
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:24 PM Revision e00a8244 (ceph): client: map file stripes to acting osds
- Old result was just wrong if any osds were down.
Signed-off-by: Sage Weil <sage@newdream.net> - 10:18 PM Revision f2c94459 (ceph): rados_sync: fix import a bit
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:11 PM Revision 8a56a74d (ceph): rados_sync: nicer error formatting
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 10:07 PM Bug #817 (Won't Fix): Implement custom propose logic for OSDMap/PGMap
- 10:07 PM Cleanup #755 (Resolved): Compartmentalize rados.cc
- 10:05 PM Bug #1028 (In Progress): segfault in OSDMap::object_locator_to_pg
- 04:48 PM Bug #1028: segfault in OSDMap::object_locator_to_pg
- added some debug checks in the code to track this one down, commit:85292b367b0e6e6d8963de32ad198482500c887f
- 09:57 PM Revision e9550b57 (ceph): common: be a little less scary in our startup warning
- - be less scary
- be helpful and tell them where to go
Signed-off-by: Sage Weil <sage@newdream.net> - 09:47 PM Revision f98475da (ceph): rados_sync: misc bugfixes
- Avoid using a temporary c_str pointer when it goes out of scope.
Complain if XATTR_FULLNAME is empty, as well as mis... - 09:37 PM Revision 6b6617ef (ceph): rados_sync: nicer printouts
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:31 PM Revision a4936a5e (ceph): man: mkcephfs: small fixes
- Signed-off-by: Sage Weil <sage@newdream.net>
- 09:28 PM Revision 73facc19 (ceph): rados_sync: do_export: fix diff'ing
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:22 PM Revision 6c9698b4 (ceph): rados_sync: more improvements to xattr_test
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 09:21 PM Revision d74ecc26 (ceph): mkcephfs.man: update man page
- Not beautiful, but at least it's accurate.
Signed-off-by: Sage Weil <sage@newdream.net> - 09:13 PM Revision 415db79d (ceph): rados_sync: Add xattr_test
- Add xattr_test to test if the underlying filesystem supports xattrs
Fix BackedUpObject::get_fs_path a bit.
Signed-o... - 08:43 PM Revision fb84e6b2 (ceph): Merge branch 'stable'
- 08:42 PM Revision 48d94f6e (ceph): osd: used fixed size types for fiemap/mapext/sparseread encoding
- The client expects <uint64_t,uint64_t>, so this breaks on any 32-bit osd.
Signed-off-by: Sage Weil <sage@newdream.net> - 08:15 PM Revision 739ab87d (ceph): cfuse: don't -d
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:15 PM Revision f757f134 (ceph): mds: don't journal src when not auth on rename
- If we aren't auth for the srcdn, don't put it in the journal, or else we'll
barf on journal replay.
This fixes a cra... - 08:15 PM Revision 4fbd6c23 (ceph): mds: assert auth on journal_cow_dentry
- We have no business doing cow (or journaling) non-auth dentries.
Signed-off-by: Sage Weil <sage@newdream.net> - 07:25 PM Revision 621af5e3 (ceph): PG: choose_acting needs the value of the osd, not its index
- This caused two osds to keep flipping the acting set between [2] and
[0,2] when osd.0 was far behind and needed a bac... - 07:10 PM Bug #1074 (Resolved): rados_sync: implement --delete
- Implement a --delete flag that will get rid of objects in the dest that aren't in the source.
Should be pretty sim... - 07:06 PM Revision 2315a64e (ceph): Create rados_sync tool to back up rados pools
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 06:10 PM Revision 0edbc75a (ceph): stop.sh: Avoid bashisms.
- I have a habit of running "sh -x stop.sh" whenever it seems
to fail, and that runs it with dash, not bash. Since it
d... - 06:07 PM Revision 6ee6261c (ceph): stop.sh: Whitespace.
- Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>
- 05:42 PM Bug #1073 (Resolved): osd: failed assert: info.last_complete == info.last_update
- I got this crash on several osds:...
- 04:26 PM Revision 6c366c06 (ceph): osd: fix include in cosd.cc too
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 04:26 PM Revision 88875fba (ceph): osd: rearrange #includes to get our assert
- Make sure we include boostchart headers before our common/assert.h so that
ours clobbers theirs. Otherwise the gener... - 04:10 PM Bug #1072 (Won't Fix): cfuse mount fails when monitors have incorrect time synchronization
- I've run into this a few times, but this time was the most obvious.
client debugging:
2011-05-06 21:40:11.448803 7f... - 03:07 PM Linux kernel client Bug #1071: rbd: mkfs.ext4 doesn't complete (but mke2fs -j does)
- bisected this down to
fatty:ceph-client 03:09 PM $ git bisect bad
77f38e0eeac290827f41fd2215ab82546b8f73b8 is t... - 09:51 AM Linux kernel client Bug #1071 (Resolved): rbd: mkfs.ext4 doesn't complete (but mke2fs -j does)
- root@uml:~# mke2fs /dev/rbd0
mke2fs 1.41.12 (17-May-2010)
Filesystem label=
OS type: Linux
Block size=4096 (log=... - 02:14 PM CephFS Bug #1063 (In Progress): dbench breaks if MDS and client times aren't synced
- Ran this with client debugging enabled (job 573). Not certain this is the problem, but it looks like the problem is t...
- 02:04 PM Bug #1066 (Resolved): update mkcephfs man page
- 01:45 PM rgw Bug #1059: RGW consistency issues
- IIRC we settled on:
- if pool is not in the osdmap, check for the bucket object. if that also doesn't exist, retu... - 12:36 PM rgw Bug #1059: RGW consistency issues
- Other than one test that failed due to misconfiguration, the rest failed due to the async bucket creation. We should ...
- 10:58 AM CephFS Bug #1041 (In Progress): standby-replay fails on multi-mds fsstress journals
- The problem is that the journal (for mds0) refers to mds1's stray directory. It's replaying a rename operation, where...
- 09:40 AM rbd Bug #1070 (Rejected): krbd: ^C doesn't work
- root@uml:~# dd if=/dev/zero of=/dev/rbd0 bs=1M count=100
^C
^C^C^C
100+0 records in
100+0 records out
10... - 04:42 AM CephFS Bug #1047: mds: crash on anchor table query
- I alredy many times create cluster "from scratch", but log still available.
Full mds log from start after previous...
05/05/2011
- 11:36 PM Revision 2ab3062b (ceph): PG: strays don't peer (only the primary does)
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 11:35 PM Revision 76febd2d (ceph): PG: clear peering flag when we transition from Peering to Reset
- This is safe because the only other transitions are to Pending (still
peering) and Active, which clears the state its... - 11:27 PM Revision d6ebf600 (ceph): mon: fix multi-osd 'ceph in|out N..' operations
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:24 PM Revision 3ee41ab8 (ceph): ceph_argparse: ignore dashes after equals sign
- When parsing argv, ignore dashes after equals signs. This is so that
things like --log-file=/tmp/foo-file will work c... - 11:17 PM Revision 57bc9b3f (ceph): Merge branch 'wip_pg_refactor'
- 11:17 PM Revision 3b92d408 (ceph): Merge branch 'stable'
- 11:12 PM Revision dc9be200 (ceph): osdmap: fix temp osd pg mapping
- If you feed in a raw pg (full precision) you should get the same mapping
out as when you plug in the effective/reduce... - 10:24 PM Revision 58ea8358 (ceph): Merge branch 'stable'
- Conflicts:
src/mds/Resetter.cc - 10:15 PM Revision 5c520fef (ceph): mon: do not stop mds0 unless all other nodes have also stopped
- If we are the root node or the tableserver, we have to shut down last.
(And even then, if we have client sessions, we... - 09:36 PM Revision 6ac55729 (ceph): PG: handle MOSDPGLog messages in Active state.
- We previously didn't, but we definitely need to as we can go
active while waiting on strays to send in logs to recove... - 09:07 PM Revision 79b1a101 (ceph): ceph_crypto: Fix ceph::crypto::init mutex for NSS.
- Even after ceph::crypto::shutdown, the library looked
already initialized; this broke the ForkDeathTest.MD5
in "make ... - 08:42 PM Revision 44900d4c (ceph): v0.27.1
- 08:38 PM Revision 27a48c1c (ceph): ceph::crypto::init: add PTHREAD_MUTEX_INITIALIZER
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:35 PM Revision 7d12d180 (ceph): mds: fix --reset-journal
- Don't fork. (Already fixed in master branch by the start_with_nonce
refactor, so this is just for 0.27.1.)
Signed-o... - 08:27 PM Revision 86064491 (ceph): test-obsync.py: allow better nonce,better printout
- Support deterministic nonces. Better debug printouts.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 08:27 PM Revision 28ccdf42 (ceph): obsync: remove misguided strip_owner
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 08:07 PM Revision 2ca61c92 (ceph): osdmap: small cleanup
- Signed-off-by: Sage Weil <sage@newdream.net>
- 08:07 PM Revision 8460cffa (ceph): libceph: initialize keyring after conf is parsed
- Init the keyring in ceph_mount, after the user has a change to load a conf
file or parse command line args.
Signed-o... - 08:07 PM Revision e6dfcfda (ceph): fix some shadowing arguments
- Small subset of what -Wshadow turns up.
Signed-off-by: Sage Weil <sage@newdream.net> - 07:44 PM Revision 921d4b3d (ceph): crypto: fix non-reentrancy of ceph::crypto::init
- This could be called multiple times from common_preinit.
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com> - 06:39 PM Revision 429bf1fe (ceph): test-obsync: use s3-tests configuration file
- Use the same configuration file format as s3-tests. This should make it
easier for people to run obsync tests!
Signe... - 06:20 PM Revision b8055b45 (ceph): osd: fix GetInfo down check
- The PgPriorSet::down set can have lots of stuff it in without it affecting
peering completion. We just need to look ... - 05:55 PM Revision db4026d2 (ceph): osd: some comment, debug output cleanup
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:54 PM Revision a284e205 (ceph): osd: reg_last_scrub on pg load
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 05:22 PM Revision 2b1c61e4 (ceph): obsync: remove omit_owner for acls
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:19 PM Revision 6d8ddbc9 (ceph): obsync: Fix ACL parsing yet again
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:19 PM Revision b39bb462 (ceph): boto_tool: add getacl, setacl
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:19 PM Revision 24f641f8 (ceph): obsync: be more verbose about set_xml_acl errors
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:19 PM Revision 8964359c (ceph): obsync: implement --force option
- Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
- 05:09 PM Revision 5d77c543 (ceph): osd: GetLog, GetMissing cleanups, fixes
- Added missing continue in GetMissing ctor.
Signed-off-by: Sage Weil <sage.weil@dreamhost.com> - 05:03 PM Revision fe988b11 (ceph): osd: only transition to Pending from GetInfo cons
- No need to do so in get_infos(). The need_up_thru requirement will only
get better as last_epoch_started floor moves... - 04:39 PM Cleanup #1009 (Resolved): osd: push pg message handling code into ReplicatedPG
- 04:38 PM CephFS Bug #1041: standby-replay fails on multi-mds fsstress journals
- I've got a log in kai:~gregf/logs/fsstress/standby-replay
- 04:37 PM Bug #910 (Resolved): Multi-MDS Ceph does not pass fsstress
- Haven't seen any new issues!
- 04:36 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- I'm unable to reproduce this on my own machine, and after looking through the mds logs from autotest everything looks...
- 02:41 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Job 560 has spent 1.5 hours in cleanup now, aborting.
14:40:26 DEBUG| [stdout] 2 19 0.00 MB/sec cle... - 12:55 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- My bad, the cleanup phase starts at 600 seconds, so kclient only had a few seconds of cleanup.
The cfuse re-run is... - 12:50 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Job 555 broke, here's a re-run: http://autotest.ceph.newdream.net/afe/#tab_id=view_job&object_id=556
And that conf... - 09:14 AM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- Note that the test I ran was on cfuse (most likely because I had kclient trouble, and wanted to isolate that out).
... - 04:30 PM Revision d2865501 (ceph): osd: consistent naming
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:54 PM Revision 04e343c9 (ceph): osd: handle notify+info explicitly in GetInfo state
- This fixes a few things:
- do not proceed past GetInfo if there are down osds. ever.
- if we get a new info that m... - 03:28 PM CephFS Bug #1047: mds: crash on anchor table query
- BTW, if you know what directory the hard link(s) in this directory reference (it'll be the first time the linked file...
- 03:27 PM CephFS Bug #1047: mds: crash on anchor table query
- It's the crash above, I take it?
Can you include a full mds log that ends with an 'ls' and crash?
It sounds lik... - 03:24 PM Bug #1048 (Resolved): MDS never finishes "stopping"
- commit:5c520fefce50c29fbf05a174a37346d8f728c3ef
i suspect if we have no sessions there might be problems, but that... - 03:18 PM Revision ade1fcb7 (ceph): osd: statechart whitespace
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:14 PM Revision 565de261 (ceph): osd: initialize pg state event counters
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 03:12 PM Revision c24859f1 (ceph): osd: fix GetInfo querying
- Don't query for info we already have, or have already requested. Remove
unneeded helper so that this is simpler and ... - 03:11 PM Revision 1fbefb32 (ceph): osd: handle event notify/info/log from Initial
- We shouldn't post a creation event and jump into peering/stray based on
pg creation when we are about to process more... - 02:31 PM Feature #1069 (Resolved): rados: support attrs in import/export
- store these as xattrs on the files... user.rados.* i guess!
- 02:30 PM Feature #1068 (Resolved): rados: incremental import/export
- Skip objects when size/mtime and attrs match on source and destination. Or sync always. (Make it an option.)
We ... - 02:01 PM Bug #1058 (Resolved): pg_refactor: OSD crash when marking several out
- Haven't seen a recurrence of this.
- 02:01 PM Bug #1062 (Resolved): pg_refactor: active primary is getting MPGLog
- Fixed my bug with commit:6ac5572946337e8e74b5014fc299e0ee10c70d9f and I think Sage already got his, too.
- 01:35 PM CephFS Bug #1064: all mds's dies one by one after restart
- 01:22 PM CephFS Bug #1064: all mds's dies one by one after restart
- Hi Sergey,
Can you attach the full mds log for journal replay? (probably need to gzip, it'll be big!)
Also, ca... - 03:52 AM CephFS Bug #1064 (Can't reproduce): all mds's dies one by one after restart
- At first all cfuse client hungs while trying access files on mounted ceph. I restart all nodes with ceph -a stop && c...
- 01:18 PM CephFS Bug #1065 (Resolved): mds: can't authenticate with --reset-journal
- commit:7d12d1807fde72d6d87c732a0bda6b33bb4d8f7c
- 09:38 AM CephFS Bug #1065 (Resolved): mds: can't authenticate with --reset-journal
- 08:54:46 AM) Tsipa: trollface: when i'm running cmds -i 0 -c /etc/ceph/ceph.conf --reset-journal 0 i get messages lik...
- 01:16 PM Feature #1067 (Resolved): mkcephfs: magically group osds on same host into subtrees in the genera...
- In theory we can look at the host field in the ceph.conf to generate a sane crushmap by default. As things stand, an...
- 11:11 AM Bug #1066 (Resolved): update mkcephfs man page
- 10:36 AM rgw Bug #1059: RGW consistency issues
- Here is my run of s3-tests. Looks like these tests failed:
test_s3.test_bucket_list_empty
test_s3.test_bucket_cre... - 10:22 AM Bug #1057 (Resolved): obsync: implement --force
- implemented by commit:8964359c9d0a56e72b64c81a2ceda1ec3c29c5be
- 12:25 AM Revision d6347392 (ceph): rgw: don't output optional DisplayName if not set
- Don't output optional DisplayName if not set. Be a little more explicit
about zeroing other optional fields during xm...
05/04/2011
- 11:44 PM Revision b5cb4c5e (ceph): osd: debug handle_*
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:19 PM Revision a544bda7 (ceph): rgw: don't assume XML is valid
- XML parsing functions in RGW now return a bool, indicating whether they
were able to get the fields they needed to ou... - 11:17 PM Revision 4d10126b (ceph): OSD: fill in rctx properly for pg->handle_create in get_or_create_pg
- Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
- 11:10 PM Revision 4d661999 (ceph): osd: rename states to reflect nesting; fix enter/exit msgs
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 11:10 PM Revision 9ebcd365 (ceph): osd: fix min_time in state stats
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 09:58 PM Revision d2c245c4 (ceph): osd: first pass at pg peering stats
- The numbers are a bit off it seems. Also lots of potential for cleanup
here. But it (basically) works!
Signed-off-... - 09:36 PM CephFS Bug #1063: dbench breaks if MDS and client times aren't synced
- this is probably a kclient thing.. testing against latest for-linus
- 04:56 PM CephFS Bug #1063 (Can't reproduce): dbench breaks if MDS and client times aren't synced
- http://autotest.ceph.newdream.net/afe/#tab_id=view_job&object_id=554
one mds, one osd, cfuse
dbench never compl... - 09:30 PM Revision 228e857e (ceph): OSD: assert contents exist when erasing from last_scrub_map.
- Insert PG into last_scrub_map on creation so that this doesn't
break right away.
Signed-off-by: Greg Farnum <gregory... - 09:12 PM Revision a347d180 (ceph): osd: use const char * state names
- Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
- 08:52 PM Revision ca7a3724 (ceph): PG: proc_replica_info, oinfo not info
- The method param info shadowed PG::info.
- 08:05 PM Revision db8f96c9 (ceph): osd: move directly to Reset state on pg load
- Add Initial -> Reset transition on pg load. This avoids doing any
activation-type stuff (like sending messages) befo... - 07:23 PM Revision d622a6fb (ceph): OSD: start PG state machine when loading pre-existing PGs
- This caused a crash when restarting a killed OSD because the Initial
state was receiving the ActMap event.
Signed-of... - 07:03 PM rgw Bug #1059: RGW consistency issues
- Can you specify which tests are failing? I haven't seen this happening, but I was probably running it against a diffe...
- 04:40 PM rgw Bug #1059: RGW consistency issues
- I think this issue is causing a lot of the s3-tests failures I'm seeing.
- 03:27 PM rgw Bug #1059 (Resolved): RGW consistency issues
- RGW is supposed to implement read-after-write consistency, but it often does not....
- 06:36 PM Revision c3fe4e3d (ceph): PG: ReplicaActive must repond to requests from discover_all_missing
- If the peer does not yet have the pg during GetMissing, there won't be
a peer_missing entry for that peer. In that c... - 05:50 PM Revision c9859c2f (ceph): uclient: only try to update caps on the auth MDS.
- Previously we would send updates on things like the max_size we
wanted to the first MDS in our list, which was bad if... - 05:50 PM Revision d417fb0b (ceph): uclient: fill in InodeCap::wanted member on send_cap.
- Previously it was never used, which I think led to spamming the MDS
with cap messages!
Hopefully this won't reveal a... - 05:30 PM Bug #1053: rgw XML parsing exploits and flaws
- that should read:
commit:a544bda7577321c4d6ecf7664a9363180984da56 and commit:d6347392634678b93f510f98d3d42407d05a956c - 05:29 PM Bug #1053 (Resolved): rgw XML parsing exploits and flaws
- I was afraid that there were more mistakes like #1055 in the code, but it looks like the other XML parsing stuff isn'...
- 10:49 AM Bug #1053 (Resolved): rgw XML parsing exploits and flaws
- RGW must never segfault on bad network input. It should handle optional XML fields.
- 05:28 PM Revision 9c192e58 (ceph): PG: collapse crashed transitions to happen on any unexpected event
- Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
- 05:27 PM Revision 8d301d29 (ceph): PG: use a state_name member instead of overriding get_state_name
- Also add debugging to each state constructor. Since dout uses
the recovery machine context, anything using it in the ... - 04:57 PM Tasks #1055 (Resolved): RGW segfaults if the Owner field is not set in an ACL (subtask)
- fixed by a544bda7577321c4d6ecf7664a9363180984da56
- 10:53 AM Tasks #1055 (Resolved): RGW segfaults if the Owner field is not set in an ACL (subtask)
- The following code explains the situation quite well, I think:...
- 04:52 PM Bug #1062: pg_refactor: active primary is getting MPGLog
- I have a slightly different crash, but i think it's the same problem. see this:...
- 04:17 PM Bug #1062 (Resolved): pg_refactor: active primary is getting MPGLog
- Logs are in kai:~gregf/logs/pg_refactor/active_gets_pglog
Same steps to reproduce as #1058.
Causes a crash in t... - 04:17 PM Bug #1058: pg_refactor: OSD crash when marking several out
- So the problem is that OSD::get_or_create_pg calls pg->handle_create using an rctx that is empty except for an empty ...
- 02:09 PM Bug #1058 (Resolved): pg_refactor: OSD crash when marking several out
- To reproduce:
1) Start up with 4 OSDs (using vstart)
2) mark two of the OSDs out
3) Wait a bit
4) Notice that the... - 11:48 AM Bug #1057 (Resolved): obsync: implement --force
- Implement --force, which causes all copies to happen, even if the source and destination appear to be in sync.
- 11:46 AM Bug #1056 (Won't Fix): obsync: add warning when source owner is not the same as dest owner (after...
- when -v is used, we should spit out a little message when we change the Owner during a source->dest copy
- 10:54 AM Bug #1054 (Duplicate): RGW segfaults if the Owner field is not set in an ACL
- re-created this as a subtask of 1053 (which I always wanted it to be)
Apparently you can't subtask-ize things afte... - 10:51 AM Bug #1054 (Duplicate): RGW segfaults if the Owner field is not set in an ACL
- The following code explains the situation quite well, I think:...
- 10:16 AM CephFS Tasks #1039 (Resolved): cfuse: requests max_size from non-auth MDS
- Looks good. Pushed to master in commit:c9859c2fc8005566ad6159a3d66e2ba10f82a986.
- 05:48 AM Bug #1018: error on building ceph on red hat 5.5
- I have had the same problem using 0,27 version.
The problem seems to be in libedit.
i have user libedit from RPMfor... - 12:58 AM Revision 879d6b03 (ceph): Revert "osd: simplify check for unconsumed events"
- This reverts commit ab34a3ce3e757a54816bd9b884c3f900361d4930.
It turns out that unconsumed_event supersedes checking... - 12:19 AM Revision 31f6a8cf (ceph): PG: Primary should also discard the ActMap event
- Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
- 12:03 AM Revision ab34a3ce (ceph): osd: simplify check for unconsumed events
- No need for the Crashed pseudo state.
Signed-off-by: Sage Weil <sage@newdream.net> - 12:02 AM Revision dda3261e (ceph): PG: ActMap should be dicarded if no outer state handles it
Also available in: Atom