Activity
From 02/04/2013 to 03/05/2013
03/05/2013
- 07:47 PM Feature #4356 (Fix Under Review): libcephfs: expose osd topology
- 07:46 PM Feature #4356 (Closed): libcephfs: expose osd topology
- wip-expose-topo
e8da4bf and 6b3fce1 - 07:14 PM Feature #4074 (Fix Under Review): qa: add traceless reply test to fs suite
- 07:14 PM Fix #4034 (Fix Under Review): mds: fix replayed ino creation extra_bl
- 03:47 PM Feature #4355 (New): uclient: add perfcounters
- The client currently has 3 perfcounters: average latency of replies, of processing a request, and of a file write.
... - 03:46 PM Feature #4354 (Resolved): mds: add an equivalent to the OSD OpTracker
- Like it says — we want to be able to get information about ops-in-flight and their current status in a lot of differe...
- 01:29 PM Cleanup #4166 (Resolved): ceph: simplify ceph_sync_write() page_align
- This has been committed to the ceph-client testing branch.
038832c ceph: simplify ceph_sync_write() page_align cal... - 11:12 AM Bug #4350 (Rejected): ceph-fuse: lockup from 40g loopback mkfs.ext3
- The underlying RADOS cluster in this report isn't fully healthy. I'm pretty sure that's all there is. Unless we hear ...
- 09:00 AM Bug #4350 (Rejected): ceph-fuse: lockup from 40g loopback mkfs.ext3
- ...
03/04/2013
- 02:10 PM Feature #4326 (Resolved): qa: add samba + (kclient|ceph-fuse) to suite
- 11:01 AM Documentation #3796 (Resolved): FUSE mount documentation needs some corrections for v0,56x
- Page has been updated with instructions, and a hyperlink to Cephx Configuration Reference. See http://ceph.com/docs/m...
- 10:06 AM Documentation #3796 (In Progress): FUSE mount documentation needs some corrections for v0,56x
- 10:38 AM Cleanup #4166: ceph: simplify ceph_sync_write() page_align
- This patch (2/3 below) has been posted for review, along with
a few others I include here for context. Marking for ... - 08:50 AM Cleanup #4166 (In Progress): ceph: simplify ceph_sync_write() page_align
- I'm reopening this after all.
It turns out that the original patch was fine. The only part
that was bad was due ... - 07:14 AM Feature #4277: Move built hadoop artificats to download URL
- Thanks for the info Gary! Let me do a little bit more research on how users want to obtain the artifacts. I think tha...
03/01/2013
- 04:11 PM Fix #2215: ceph-fuse does not invalidate page cache
- Which automatic tests actually run those? I'm not sure that the nightlies do so right now.
- 04:00 PM Fix #2215: ceph-fuse does not invalidate page cache
- I add the fuse_use_invalidate_cb: true option in the ceph-qa-suite to the basic and verify fs suites (in the btrfs.ya...
- 02:36 PM Feature #3819: mds: re-add snaptests to qa suite
- De-prioritizing in light of ongoing discussions.
- 02:35 PM Bug #4134: mds: request locking hang under snaptests
- De-prioritizing this for now based on our meetings and discussion yesterday.
- 02:18 PM Documentation #3113: Ceph FS Options Could Use Some Additional Information
- Any discussion of cephfs options (which really don't have much to do with mount!) should also refer to the equivalent...
- 01:45 PM Documentation #3113: Ceph FS Options Could Use Some Additional Information
- John - could you move this up your queue?
- 02:05 PM Fix #4034: mds: fix replayed ino creation extra_bl
- 01:44 PM Documentation #2206: Need a control command to gracefully shutdown an active MDS prior to planned...
- 'ceph mds fail 0'
- 11:41 AM Bug #4308 (Won't Fix): ceph-fuse crashed during blogbench test (argonaut)
- Log: ubuntu@teuthology:/a/teuthology-2013-02-28_20:00:04-regression-argonaut-master-basic/14324...
02/28/2013
- 04:51 PM Feature #4074 (In Progress): qa: add traceless reply test to fs suite
- 10:48 AM Feature #4295: mds: Actually purge deleted directories
- And Sage points out the reason it's not done so far is because fragmentation mucks it up a bit....
- 10:48 AM Feature #4295 (Resolved): mds: Actually purge deleted directories
- Right now we never actually delete a directory object when it gets unlinked in the filesystem.
1) Identify the reaso... - 09:33 AM Fix #4286: SLES 11 - cfuse: disable 'big_writes'and 'atomic_o_trunc
- Need to figure out what version of fuse added those flags.
02/27/2013
- 10:42 PM Tasks #4222 (Resolved): Add libcephfs-test.jar to ceph-test deb package
- Resolved with commit:
commit b65ca564b6ec7324ba43e4d35a2d114b4f7220ea
Author: Gary Lowell <glowell@inktank.com>
... - 11:10 AM Fix #4286 (Rejected): SLES 11 - cfuse: disable 'big_writes'and 'atomic_o_trunc
- encounted error when mounting a cfuse volume on SLES 11.
unknown option: big-writes
unknown option: atomic_o_trun...
02/26/2013
- 10:32 PM Feature #4277: Move built hadoop artificats to download URL
- The output of the gitbuilders are staged to gitbuilder.ceph.com, the release process (different set of scripts) produ...
- 02:17 PM Feature #4277 (Closed): Move built hadoop artificats to download URL
- We are at the point now where we are building some useable jar files that we'd like to reference from the Hadoop docu...
- 05:01 PM Bug #4134: mds: request locking hang under snaptests
- Oh, and actually the client is the one with the seq of 2 while the MDS only thinks it should have a seq of 1. Curious...
- 02:48 PM Bug #4134 (In Progress): mds: request locking hang under snaptests
- The request is attempting to read a snapshotted inode, but the inode is still in the LOCK_SNAP_SYNC state from when i...
- 04:32 PM Bug #4280: mds: crash on lookupsnap
- Logs and core in kai:/home/gregf/logs/4280
- 04:31 PM Bug #4280 (Resolved): mds: crash on lookupsnap
- While running qa/workunits/snaps/snaptest-snap-rm-cmp.sh:...
- 09:36 AM Bug #4266 (Won't Fix): crush placement from crushtool mismatches observed behavior for pools othe...
- The --x is only coincidentally related to the second part of hte pgid (currently). If you want to map a specific pgi...
- 08:05 AM Bug #4266 (Won't Fix): crush placement from crushtool mismatches observed behavior for pools othe...
- Because OSDs add the pool id to the pg number to form the x passed to crush_do_rule, whereas crushtool just gets an x...
02/25/2013
- 01:51 PM Bug #4248: mds: replay does not correctly update CInode::first and ::last members
- This isn't currently journaled correctly, in addition to not setting its snapid ranges correctly. There's nothing tha...
- 11:19 AM Bug #4248 (In Progress): mds: replay does not correctly update CInode::first and ::last members
- Okay, that didn't fix the problem. The stray dir now was snapids [3, HEAD] instead of [4, HEAD], but the inode remain...
- 09:51 AM Feature #3821 (New): qa: run backuppc as part of qa suite
02/22/2013
- 03:32 PM Bug #3794 (Resolved): uclient: reports sizes wrong in some cases
- 03:10 PM Bug #3794: uclient: reports sizes wrong in some cases
- Assuming we aren't pushing this into a stable release without more testing, both the patch and backport are
Review... - 02:58 PM Bug #3794 (Fix Under Review): uclient: reports sizes wrong in some cases
- wip-statvfs
- 03:31 PM Bug #3793 (Resolved): wrong size reported in some distributions/toolchains
- 'ceph: fix statvfs fr_size' in kernel tree.
- 02:58 PM Bug #3793: wrong size reported in some distributions/toolchains
- I pushed a wip-statvfs which fixes this for ceph-fuse.
- 03:22 PM Bug #4248 (Fix Under Review): mds: replay does not correctly update CInode::first and ::last members
- wip-4248-snapid-journaling
Also pushing out a v0.57-based branch to the user to try out. - 03:17 PM Bug #4248 (Resolved): mds: replay does not correctly update CInode::first and ::last members
- This came in over the mailing list....
- 01:14 PM Bug #4035 (Rejected): Ceph doesn't recover from fault on Opensuse (cfuse tests & rbd-cli tests)
- 01:11 PM Bug #3818 (Duplicate): kclient: fsx fails in mapread
- 01:09 PM Bug #3681: kclient fsx fails nightly
- Should review entire kernel locking around truncate.
- 01:08 PM Bug #3681: kclient fsx fails nightly
- 12:45 PM Bug #2277: qa: flock test broken
- It's not an issue because these tests aren't being run any more. They should be run, though. ;)
That said, this wa... - 12:30 PM Bug #2277 (Closed): qa: flock test broken
- No longer an issue.
- 12:35 PM Bug #4241 (Duplicate): SELinux fails because it can't set xattrs
- ...
- 12:23 PM Bug #1874 (New): Running `git gc` on a bare git repository hosted by ceph results in a bus error.
02/21/2013
- 02:22 PM Bug #4220 (Resolved): MDS is inconsistent about whether layouts are allowed on the root directory
- Right, this is in the generic xattr code, not the new vxattr stuff. Sage says this is because the root inode used to ...
- 02:03 PM Bug #4220: MDS is inconsistent about whether layouts are allowed on the root directory
- If you go through the "handle_client_setdirlayout" interface (ie, cephfs tool), the MDS will let you set a layout on ...
- 10:47 AM Bug #4220 (Resolved): MDS is inconsistent about whether layouts are allowed on the root directory
- I can't seem to set layouts on the root directory with ceph-fuse. Perhaps I'm doing something else wrong, but check o...
- 01:50 PM Bug #1435 (Resolved): mds: loss of layout policies upon mds restart
- Change pushed to bobtail (commit:36ed407e0f939a9bca57c3ffc0ee5608d50ab7ed) and next (commit:6bd8781dda524f04bb56bcdac...
- 01:49 PM Cleanup #1499 (Resolved): mds: clean up directory layouts
- Okay, I tested this according to the procedure described below (I wanted to write some scripts, but the non-manual po...
- 11:28 AM Tasks #4222 (In Progress): Add libcephfs-test.jar to ceph-test deb package
- 11:26 AM Tasks #4222 (Resolved): Add libcephfs-test.jar to ceph-test deb package
- The file libcephfs-test.jar is not present in any of the deb packages.
The file ceph.spec was changed to include thi... - 10:48 AM Bug #4221 (Resolved): MDS: LogEvent::decode needs to respect mds_log_skip_corrupt_events for DECO...
- By far the most common form of corrupt event is one that's somehow the wrong size — and that hits an assert in LogEve...
02/20/2013
- 05:35 PM Bug #1435 (Fix Under Review): mds: loss of layout policies upon mds restart
- Okay, after discussing with Sage we've decided that placing the directory layouts into the inode_t::layout field is t...
- 11:36 AM Bug #1435 (Won't Fix): mds: loss of layout policies upon mds restart
- This patch to use the inode layout instead of default_layout shouldn't be helping — directory layouts aren't written ...
- 05:33 PM Feature #4215 (Resolved): mds: support setfattr on ceph.dir.layout
- Ah, it was just a lost commit! Resolved by commit:5551aa5b3b5c2e9e7006476b9cd8cc181d2c9a04, among others.
(It will p... - 02:43 PM Feature #4215: mds: support setfattr on ceph.dir.layout
- Oh, as a feature this isn't any particular priority. Just an optimization, really.
- 02:40 PM Feature #4215: mds: support setfattr on ceph.dir.layout
- Oh, and indeed, looking at the source code:...
- 02:37 PM Feature #4215 (Resolved): mds: support setfattr on ceph.dir.layout
- ...
- 04:52 PM Cleanup #1499 (Fix Under Review): mds: clean up directory layouts
- Okay, this patch is required to fix the loss of directory layouts issue (#1435) in Bobtail, so I rebased it on that b...
- 12:04 PM Bug #4213 (Resolved): mds: old_parents is never cleaned up
- The CInode::old_parents map is never cleaned up, even if all the snapshots for which a parent is valid have all been ...
- 12:01 PM Bug #4212 (Closed): mds: open_snap_parents isn't called all the times it needs to be
- Prompted by a bug report/paper-over patch from Alexandre, I had a talk with Sage about snapshots. It came out that we...
- 10:01 AM Feature #3730 (Closed): Support replication factor in Hadoop
- Woot.
- 09:43 AM Feature #4208: Add more replication pool tests for Hadoop / Ceph bindings
- Something like this added to the current custom test should do the trick?...
- 09:32 AM Feature #4208 (Rejected): Add more replication pool tests for Hadoop / Ceph bindings
- We should add a test (or several) for the scenario where the user specifies a level of replication that does not exac...
02/19/2013
- 05:48 PM Cleanup #4166: ceph: simplify ceph_sync_write() page_align
- This commit caused regressions. I pulled it out and
I'm not going to re-open it again.
- 11:55 AM Fix #4191 (Resolved): qa: mulitiple mds in nightly (non-failure case)
- 11:46 AM Feature #4190 (Resolved): qa: add mds thrashing to nightly
- 11:43 AM Feature #4002 (Resolved): mds: design fsck
- Okay, this is as done as it needed to be.
- 10:18 AM Bug #1435: mds: loss of layout policies upon mds restart
- Oh, n/m, hurray for our sysadmins fixing hasty upgrades. :)
- 10:16 AM Feature #3730: Support replication factor in Hadoop
- Yes. I should have closed this and opened a separate ticket for the tests. I'm planning to close it as soon as Joe wi...
- 10:14 AM Feature #3730 (In Progress): Support replication factor in Hadoop
- Isn't this now merged or something? :)
- 09:43 AM Bug #3935: kclient: Big directory access bugs (multiple), mixed 32- and 64-bit clients
- I was thinking the offset for readdir might be 32-bits.. but i may be wrong there.
- 03:07 AM Bug #4188 (Can't reproduce): mds crashes when cow-ing entries in formerly snapshotted dir
- I have a dir that I used to snapshot before I switched to hardlink trees. I create files and subdirs in it, then, wh...
02/18/2013
- 05:00 PM Cleanup #4166 (Resolved): ceph: simplify ceph_sync_write() page_align
- commit 29c3c8b721a96b4a82f4224527c7103e5e910b80
Author: Alex Elder <elder@inktank.com>
Date: Fri Feb 15 22:10:16 ...
02/16/2013
- 06:35 PM Bug #1435: mds: loss of layout policies upon mds restart
- Uhh, sorry, I forgot an artifact of the first patch made to the bug-fix patch. s/default_layout_()/default_layout/g ...
- 06:31 PM Bug #1435: mds: loss of layout policies upon mds restart
- The attached patch seems to fix the problem. It's clear I'm not entirely happy with the idea of testing the inode la...
- 12:17 PM Bug #1435: mds: loss of layout policies upon mds restart
- I've been looking some more into this. It seems that my hunch was correct: the loss seems to occur at cluster recove...
- 09:12 AM Cleanup #4166: ceph: simplify ceph_sync_write() page_align
- The following has been posted for review to resolve
this issue:
[PATCH] ceph: simplify ceph_sync_write() page_ali... - 08:12 AM Cleanup #4166 (Resolved): ceph: simplify ceph_sync_write() page_align
- (This work is done.)
ceph: simplify ceph_sync_write() page_align calculation
In ceph_sync_write() there is some...
02/15/2013
- 10:20 PM Feature #4161 (Resolved): MDS: add file layout to head object
- In support of #4145, we're going to need a way to get a file's layout even if we lose the containing directory. Add a...
- 05:20 PM Feature #3857 (Resolved): mds: enforce unique mds names in mdsmap
- commit:c8f090aed96d615b081f5b9aa36c74092371ae1e
- 04:43 PM Feature #3573 (Rejected): ceph-fuse: Add ioctl support
- 04:41 PM Feature #2195 (Resolved): Allow removal of last MDS if there's no filesystem
- this works now.. 'fail' and them 'rmfailed'.
- 10:42 AM Bug #4061 (Can't reproduce): mds crashed at LogEvent::decode
- Burnupi60 had some hardware issue so I had it reimaged. I reinstalled the master branch and am not able to reproduce ...
02/14/2013
- 06:57 PM Feature #4145 (Resolved): MDS: design and implement a backwards-scanning fsck
- We've discussed this a little bit, but we eventually want a mechanism that looks through all the RADOS objects in our...
- 06:55 PM Feature #4144 (New): MDS: forward scrub: do not abort a scrub if its hierarchy gets migrated
- Once scrub can survive part of its hierarchy moving, modify it so that it also can survive its containing hierarchy m...
- 06:54 PM Feature #4143 (New): MDS: forward scrub: do not abort a scrub if part of its subtree gets migrated
- Following #4142, modify the scrub and add infrastructure as needed so that a scrub continues if part of the hierarchy...
- 06:51 PM Feature #4142 (Duplicate): MDS: forward scrub: Implement cross-MDS scrubbing
- Following #4141, implement the mechanisms necessary so that an MDS can start a remote scrub in order to complete a lo...
- 06:45 PM Feature #4141 (New): MDS: forward scrub: Implement non-blocking scrub
- Once we have a blocking scrub (#4140), convert it to be non-blocking. If necessary it can block the dentry which is c...
- 06:43 PM Feature #4140 (New): MDS: forward scrub: add infrastructure to perform a blocking scrub of all au...
- Write functionality for the MDS so that on command it will do a blocking scrub of all data in a tree for which it is ...
- 06:38 PM Feature #4139 (Resolved): MDS: forward scrub: add scrub_stamp infrastructure and a function to sc...
- Add the scrub stamps necessary for the scrubbing algorithm described in our forward scrub proposal; be efficient abou...
- 06:34 PM Feature #4138 (Resolved): MDS: forward scrub: add functionality to verify disk data is consistent
- Design and implement a mechanism that, given a CInode, it will go check the actual disk state of that inode for consi...
- 06:33 PM Feature #4137 (New): MDS: Implement a forward-scrubbing mechanism.
- Design and implement a system that checks the filesystem for consistency by starting at the root node.
So far we'v... - 11:02 AM Bug #4134: mds: request locking hang under snaptests
- For reference, the hung request is client.4104:75.
- 11:00 AM Bug #4134 (Resolved): mds: request locking hang under snaptests
- Haven't diagnosed this yet, but I ran into it while testing for #3819. Logs in kai:~/logs/4134.
First step, just fig... - 11:01 AM Feature #3819: mds: re-add snaptests to qa suite
- Currently blocked by #4134, and possibly other things. Will work on that as I get time, but focusing on other stuff t...
- 09:53 AM Feature #3819 (In Progress): mds: re-add snaptests to qa suite
02/13/2013
- 03:05 PM Feature #3819: mds: re-add snaptests to qa suite
- Pushed a fix to wip-mds-snap-fix that Sage checked out and will merge it once I've run through the rest of the snapte...
- 01:50 PM Feature #3819: mds: re-add snaptests to qa suite
- Well, actually adding them is simple enough, but I first started a snaps run on vstart and it's crashing the MDS on a...
- 12:57 PM Bug #4114 (Resolved): libcephfs: segfault on
- commit:7e192077ffc77f6c2a534c32f0fdf57a7d05ee7f
- 10:48 AM Bug #4114 (Fix Under Review): libcephfs: segfault on
- Pretty minor fix to set the inited flag to false. But wip-4114 could use a review.
- 09:08 AM Bug #4114 (In Progress): libcephfs: segfault on
- 08:32 AM Bug #4114 (Resolved): libcephfs: segfault on
- ...
- 10:07 AM Feature #606 (Duplicate): mds: optionally store parent attr on file objects
- #3540. We decided it wouldn't be optional.
- 09:34 AM Bug #4115 (Rejected): packaging: broken ceph-test-dbg binaries
- Sage points out that those are just the debug symbols...
- 09:26 AM Bug #4115 (Rejected): packaging: broken ceph-test-dbg binaries
- The ceph-test-dbg package installs test binaries to /usr/lib/debug/usr/bin/, but the binaries themselves don't have e...
02/12/2013
- 03:09 PM Bug #4105: mds: fix up the Dumper
- 02:00 PM Bug #4105 (Resolved): mds: fix up the Dumper
- The messenger/objecter locking is wrong, and my quick stab at a fix resulted in lockdep warnings and things. Spend so...
- 02:40 PM Bug #4060 (Resolved): mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- commit:a04c01f6822b165bf339d41eda29fcc5fa555f53
- 02:23 PM Bug #4061: mds crashed at LogEvent::decode
- Hmm, actually this one might be different from the other. It's a client cap update event, and the event on disk claim...
- 01:59 PM Bug #4061: mds crashed at LogEvent::decode
- Probably the same Tamil, yes. This should be a little easier to debug if we get it again in the future following last...
- 11:10 AM Bug #4061: mds crashed at LogEvent::decode
- There was no load yet. I was attempting the ceph-fuse mount test after the update.
Today I have ceph version 0.... - 11:05 AM Bug #4061: mds crashed at LogEvent::decode
- looks like it is same as bug#3773
- 09:44 AM Bug #4061: mds crashed at LogEvent::decode
- Ken, what was the workload you were running on this before the crash?
- 10:41 AM Bug #733: cmds crash: mds/LogEvent.cc:88: FAILED assert(p.end())
- This is at least the same crash as #4061, although it'd be nice to get one of these with logging on the caused end in...
02/11/2013
- 10:15 PM Feature #3543 (Closed): mds: new encoding
- After an fs suite run passed all tests except for libcephfs-java (which was known bad at the branch point), iogen whi...
- 01:55 PM Bug #4061: mds crashed at LogEvent::decode
- sorry Greg, I pulled the information from ken and filed this bug. Please let me know if you need more info.
- 01:36 PM Bug #4061: mds crashed at LogEvent::decode
- IIRC I was waiting on some other info from Ken for this. Is that coming? :)
- 01:39 PM Feature #3730: Support replication factor in Hadoop
- Initial set of tests are in the Hadoop tree and working. Need to add them to Teuthology test thingy. There are now tw...
- 12:39 PM Bug #3935: kclient: Big directory access bugs (multiple), mixed 32- and 64-bit clients
- The only way I can think of that a 32-bit client would be different is in the inode assignment; could it be running i...
- 05:37 AM Bug #3935: kclient: Big directory access bugs (multiple), mixed 32- and 64-bit clients
- 12:00 AM Bug #3935: kclient: Big directory access bugs (multiple), mixed 32- and 64-bit clients
- Tried with updated to 0.56.2. Found no troubles, but actually environment changed, since I removed 32-bit kernel clie...
- 10:49 AM Feature #4073: qa: add message delay injection to test suite
- I'm rewording the title, because its similar to #3570, which has tests already in the marginal suite. The purpose of...
- 10:00 AM Feature #4073 (Resolved): qa: add message delay injection to test suite
- 10:35 AM Bug #4068 (Closed): libcephfs: if client->init() fails, shutdown() erroneously calls client->shut...
- Commit 133295ed001a950e3296f4e88a916ab2405be0cc resolves this issue. The failure case no longer throws an assert but ...
- 10:00 AM Bug #4068: libcephfs: if client->init() fails, shutdown() erroneously calls client->shutdown(), r...
- Reviewed by Sage and Slang.
- 09:34 AM Bug #4068 (Fix Under Review): libcephfs: if client->init() fails, shutdown() erroneously calls cl...
- Josh - can you please review Joe's wip branch?
- 10:01 AM Feature #4074 (Resolved): qa: add traceless reply test to fs suite
- 09:48 AM Feature #4002 (In Progress): mds: design fsck
- 09:38 AM Bug #4035 (Need More Info): Ceph doesn't recover from fault on Opensuse (cfuse tests & rbd-cli te...
- the fault message itself is nothing to worry about; just a socket error that we normally recover from. can you clari...
02/09/2013
- 06:51 PM Bug #4068: libcephfs: if client->init() fails, shutdown() erroneously calls client->shutdown(), r...
- branch wip-4068-buck has one possible fix for this issue (but maybe not the cleanest option).
- 06:27 PM Bug #4068: libcephfs: if client->init() fails, shutdown() erroneously calls client->shutdown(), r...
- title should be "erroneously calls" and not the inverse.
- 06:24 PM Bug #4068 (Closed): libcephfs: if client->init() fails, shutdown() erroneously calls client->shut...
- In src/libcephfs.cc mount() function:
if the client->init() call fails, shutdown() gets called. Assuming that the ...
02/08/2013
- 05:21 PM Bug #4060: mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- Also, I've tested this fix with a basic script:
set -e
mnt=$1
touch ${mnt}/foo.$$
rados mkpool foo.$$
poolid=$... - 05:20 PM Bug #4060 (Fix Under Review): mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- Pushed a proposed fix to wip-4060. Needs review.
- 03:18 PM Bug #4060 (In Progress): mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- Even with add_data_pool I get EINVAL, so I'm reopening this. I've verified that (as before), the osdmap on the mds ...
- 02:57 PM Bug #4060 (Rejected): mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- You need to add it to the MDSMap first ("ceph mds add datapool [x]" or something), and the server does at least wait ...
- 02:53 PM Bug #4060 (Resolved): mds: vxattr ceph.file.layout.pool doesn't check latest osdmap
- touch foo
create pool d2
setfattr -n ceph.file.layout.pool -v d2 foo
> returns EINVAL
The problem is that the p... - 03:42 PM Bug #4061 (Can't reproduce): mds crashed at LogEvent::decode
- hit this on burnupi60, when upgrading from ceph v0.56-598-gb970d05 to 0.56.2-12-gcc16791 on 4 feb and it seems to be ...
- 01:58 PM Bug #4044 (Rejected): replay failure between MDS and client
- Never mind, this turned out to be another encoding issue. Phew!
- 01:22 PM Feature #3543: mds: new encoding
- Okay, that was an easy bug to fix. Hurray!
(The LogEvent encoding was off a little bit.)
Running it through anoth...
02/07/2013
- 02:14 PM Bug #4044 (Rejected): replay failure between MDS and client
- While testing #3543 (but that shouldn't be related to this issue), I restarted the MDS and ran into a case where the ...
- 02:11 PM Feature #3543: mds: new encoding
- 3Still haven't gotten in on teuthology (soon!), but I did some local upgrade testing. I was able to upgrade from mast...
- 10:09 AM Cleanup #1499: mds: clean up directory layouts
- I've rebased on top of the wip-mds-encode-rebased branch as wip-1499-mds-layouts, although I notice it's failing some...
- 10:09 AM Bug #1435: mds: loss of layout policies upon mds restart
- I've been totally unable to come up with a scenario for how this could happen via code inspection, so I think I'm jus...
02/06/2013
- 04:48 PM Bug #4038 (Resolved): ceph-fuse: various hangs
- He says it fixed the problem, and it's in master now. (commit: 46d7dbd3472f26926c6d048bfc3c150074bfd283)
- 04:32 PM Bug #4038: ceph-fuse: various hangs
- There's a shortcut return in CInode::_flush() that wasn't setting the new completion to done (when called from _fsync...
- 04:01 PM Bug #4038 (Resolved): ceph-fuse: various hangs
- ...
- 01:38 PM Feature #3626 (Resolved): mds: debug mode to generate traceless replies to clients
- 01:38 PM Feature #3626: mds: debug mode to generate traceless replies to clients
- Merged into master in commit:08b82b3ef6b43283e35fd4e56eb5c78651345bea.
- 01:26 PM Feature #3626 (Fix Under Review): mds: debug mode to generate traceless replies to clients
- wip-4036 (commit:4ebba50a15584c89e0c5e4c6e48618055ceb96d8). Testing it now with pjd on a vstart cluster with no trace...
- 11:01 AM Bug #4035 (Rejected): Ceph doesn't recover from fault on Opensuse (cfuse tests & rbd-cli tests)
- I'm not sure if this is exclusive to fs but on an opensuse, single node cluster, when running cfuse and rbd tests a f...
- 10:08 AM Fix #4034 (Resolved): mds: fix replayed ino creation extra_bl
- I haven't tested this, but I noticed during code inspection for other things that I believe all our recent fixes for ...
02/05/2013
- 04:22 PM Feature #3626 (In Progress): mds: debug mode to generate traceless replies to clients
- Server::set_trace_dist() sets several things on the reply:
*snapbl
*head.is_dentry
*head.is_target
*trace_bl
H... - 02:35 PM Bug #1435: mds: loss of layout policies upon mds restart
- Wait, never mind. Too excited and didn't look closely enough at the projected node struct! :)
- 01:56 PM Bug #4023 (New): kclient: d_revalidate is abusing d_parent
- See Viro's email to linux-fsdevel, http://marc.info/?l=linux-fsdevel&m=135968126020360&w=2 .
We probably need t... - 01:52 PM Bug #2753 (Resolved): Writes to mounted Ceph FS fail silently if client has no write capability o...
- wip-2753-fsync errors merged and pushed in commit:b3ffc718c93b7daa75841778b5d50ea3bc5fcc53 and fsync works properly o...
- 01:47 PM Feature #4022 (New): client: qa: test non-cached operation (force sync mode)
- Right now it's possible to run the client without going through the cacher. This isn't tested at all right now. It's ...
02/04/2013
- 04:48 PM Bug #2753 (Fix Under Review): Writes to mounted Ceph FS fail silently if client has no write capa...
- Okay, I've checked that the kernel client deals correctly with an fsync (it'll return EPERM). The client branch wip-2...
- 03:49 PM Feature #3540: mds: maintain per-file backpointers on first file object
- Comments on Github for this.
- 02:21 PM Feature #3540: mds: maintain per-file backpointers on first file object
- I've pushed the additional changes for rename and sentinel to the wip-bt2 branch. Those bits are still untested, but...
- 10:45 AM Feature #3540 (In Progress): mds: maintain per-file backpointers on first file object
- The initial review happened last week; Sam has some updates for the rename and sentinel object infrastructure now but...
- 02:58 PM Cleanup #1499: mds: clean up directory layouts
- there is an old branch that does this. note that this code changed with the wip-vxattrs work, so the rebase needs to ...
- 02:34 PM Feature #3953: kclient: get/set layout via virtual xattrs
- in testing branch, passing tests. yay!
- 02:33 PM Feature #3953 (Resolved): kclient: get/set layout via virtual xattrs
- 12:36 PM Feature #4002 (Resolved): mds: design fsck
- 10:35 AM Feature #3242 (New): samba: push plugin upstream
- I believe we decided to hold off on putting more effort into this.
- 10:34 AM Feature #3542 (Duplicate): mds: migration path for existing anchors, anchortables, etc.
- Closing in favor of #4000 and #4001.
- 10:33 AM Feature #4001 (Resolved): Implement the migration path from using the AnchorTable to using lookup...
- Actually do whatever #4000 specifies.
- 10:31 AM Feature #4000 (Resolved): Design a migration path from using the AnchorTable to using lookup-by-ino
- We're currently engaged in work to do lookup-by-ino when we get an ino we don't recognize. However, any old installs ...
- 10:27 AM Feature #3999 (Resolved): update CDir encoding
- Either following or as part of #2177, we should update CDir on-disk encoding (and possibly wire encoding) to be versi...
- 10:25 AM Cleanup #3998 (Resolved): mds: split up mdstypes
- Right now we have mdstypes and it contains both MDS-exclusive and client-shared structs. Split it up into "metadata_t...
- 10:03 AM Feature #3543: mds: new encoding
- Got an early review from Sage; now waiting for a merge review and for test results from the FS suites, which are dela...
- 09:59 AM Bug #3951 (Resolved): ceph-fuse: permissions error on create
- Fixed in master commit:cf7c3f7d3fc7b8dc3a08a4fbe4ca1c10f2cb6054 and tested that it solves the problem.
- 09:45 AM Bug #3935 (Need More Info): kclient: Big directory access bugs (multiple), mixed 32- and 64-bit c...
Also available in: Atom