Commit Graph

58601 Commits

Author SHA1 Message Date
Jeff Layton
23956dea83 client: don't take extra target inode reference in ll_link
For the life of me, I can't figure out where this reference is ever put.
We usually take a reference like this when there is an Inode **
parameter that we'll return to the caller. ll_link doesn't have one of
those, so as best I can tell this reference is just leaked.

Zheng however says that FUSE will eventually put this reference via
ll_forget. I still don't quite get how that works, but the other callers
clearly do not handle this correctly. Change the code to only make the
fuse code take this extra reference.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
2016-10-15 19:31:46 -04:00
Sage Weil
61310d41f3 Merge pull request #11422 from xiexingguo/xxg-wip-bluestore-1011
os/bluestore: fix race condtion during blob spliting

Reviewed-by: Sage Weil <sage@redhat.com>
2016-10-11 17:22:16 -05:00
Sage Weil
dbd22b7e46 Merge pull request #11394 from liewegas/wip-bluestore-shared-blob-intrusive
os/bluestore: use std::unordered_map for SharedBlob lookup
2016-10-11 13:25:32 -05:00
Kefu Chai
4f40fca56a Merge pull request #11233 from zealoussnow/wip-0927
doc: fix start development cluster operation in index.rst

Reviewed-by: Kefu Chai <kchai@redhat.com>
2016-10-11 23:48:21 +08:00
Sage Weil
7260166da2 os/bluestore: use std::unordered_map for SharedBlob lookup
Many blobs aren't shared.  Save 8 bytes per SharedBlob by using a normal
unordered_map instead of instrusive::set.

More importantly, perhaps, it avoids us having to tune the intrusive
unordered_set size manually.  std::unordered_map does this automatically
for you, but the intrusive one does not.  And it's unclear how to
statically size it given that it's a per-collection structure and we have
no idea how many objects we'll have, how many blobs per object, and how
many objects will be cloned.

Signed-off-by: Sage Weil <sage@redhat.com>
2016-10-11 11:10:51 -04:00
Kefu Chai
bee4218fb5 Merge pull request #11183 from liewegas/wip-msgr-features
msg: make loopback Connection feature accurate all the time

Reviewed-by: Kefu Chai <kchai@redhat.com>
2016-10-11 23:01:10 +08:00
xiexingguo
d410573fc0 os/bluestore: improve ExtentMap::reshard() readability a little
As ep, sp and esp can be a little confusing.

Signed-off-by: xiexingguo <xie.xingguo@zte.com.cn>
2016-10-11 22:36:04 +08:00
Sage Weil
b1853bfe98 Merge pull request #11427 from ifed01/wip-bluestore-fix-test
test/store_test: fix errors on the whole test suite run caused by the…

Reviewed-by: Sage Weil <sage@redhat.com>
2016-10-11 09:17:19 -05:00
xiexingguo
4e33747f41 os/bluestore: add sanity check to shard offset when updating
Signed-off-by: xiexingguo <xie.xingguo@zte.com.cn>
2016-10-11 22:16:34 +08:00
xiexingguo
d300929ae9 os/bluestore: fix race conditon during blob spliting
Signed-off-by: xiexingguo <xie.xingguo@zte.com.cn>
2016-10-11 22:14:22 +08:00
Igor Fedotov
776219ef6d test/store_test: fix errors on the whole test suite run caused by the lack of properites restore in some test cases
Signed-off-by: Igor Fedotov <ifedotov@mirantis.com>
2016-10-11 14:11:51 +00:00
xie xingguo
45b257013e os/bluestore: don't scan omap kvpair during fsck()
Because it does not check anything, and removing it can
speed up the process of fsck()

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
2016-10-11 22:10:54 +08:00
xie xingguo
13d0f4fec3 os/bluestore: kiill dead member from bluestore_shared_blob_t
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
2016-10-11 22:10:53 +08:00
John Spray
701d9e1e63 Merge pull request #11150 from renhwztetecs/renhw-wip-mds-server-open
mds/server: merge the snapshot request judgment

Reviewed-by: John Spray <john.spray@redhat.com>
2016-10-11 12:27:21 +02:00
John Spray
9c2dac156f Merge pull request #11209 from david-z/wip-clear-dup-logic-mdsmonitor
mon: clear duplicated logic in MDSMonitor

Reviewed-by: John Spray <john.spray@redhat.com>
2016-10-11 12:26:56 +02:00
John Spray
d9f54fbfe1 Merge pull request #11358 from stiopaa1/mds_sessionmaph_unneededCount
mds/SessionMap.h: remove unneeded use of count

Reviewed-by: John Spray <john.spray@redhat.com>
2016-10-11 12:26:03 +02:00
John Spray
4d9bce4923 Merge pull request #11362 from batrick/i17531
mds: respawn using /proc/self/exe

Reviewed-by: John Spray <john.spray@redhat.com>
2016-10-11 12:24:10 +02:00
John Spray
8ce51e5db5 Merge pull request #11359 from batrick/i17276
client: add pid to metadata

Reviewed-by: John Spray <john.spray@redhat.com>
2016-10-11 12:23:50 +02:00
Mykola Golub
47447bfba2 Merge pull request #11395 from dillaman/wip-17549
librbd: ignore notify errors on missing image header

Reviewed-by: Mykola Golub <mgolub@mirantis.com>
2016-10-11 10:55:55 +03:00
Kefu Chai
577ad3dff6 Merge pull request #11404 from bassamtabbara/wip-fix-yasm-system-check
cmake: Fix for cross compiling

Reviewed-by: Kefu Chai <kchai@redhat.com>
2016-10-11 15:29:16 +08:00
Mykola Golub
4cc823ef56 Merge pull request #11356 from dillaman/wip-17528
librbd: update internals to use optional separate data pool

Reviewed-by: Mykola Golub <mgolub@mirantis.com>
2016-10-11 08:58:45 +03:00
Gregory Farnum
42126ca6de Merge pull request #11297 from stiopaa1/mds_sessionmap_avoidCopyingPlusConst
mds/SessionMap.cc: avoid copying and add const

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
2016-10-10 15:47:15 -06:00
Gregory Farnum
1821c14c09 Merge pull request #11283 from jcsp/wip-print-legacy-client-fscid
mds: include legacy client fsid in FSMap print

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
2016-10-10 15:46:55 -06:00
Bassam Tabbara
4c5bd5d287 cmake: Fix for cross compiling
The check for yasm tool was calling uname -m which will not
work when cross compiling. Use CMAKE_SYSTEM_PROCESSOR instead.

Signed-off-by: Bassam Tabbara <bassam.tabbara@quantum.com>
2016-10-10 14:41:55 -07:00
David Zafman
3a2be4874f Merge pull request #11206 from dzafman/wip-16474
osd: Remove extra call to reg_next_scrub() during splits

Reviewed-by: Samuel Just <sjust@redhat.com>
2016-10-10 13:55:16 -07:00
Loic Dachary
ed82c02c54 Merge pull request #11399 from theanalyst/wip-release-notes
scripts/release-notes: allow title guesses from gh tags & description update

Reviewed-by: Loic Dachary <ldachary@redhat.com>
2016-10-10 22:17:18 +02:00
Abhishek Lekshmanan
4d5cb5c852 build/ops: ceph-release-notes guess pr title based on gh tags
We add a switch --use-tags which helps guess the component of the pr
based on gh tags, still needs work, but provides better estimates when
looking at really large release notes where guessing each pr would be
too much work

Signed-off-by: Abhishek Lekshmanan <alekshmanan@suse.com>
2016-10-10 21:59:56 +02:00
Abhishek Lekshmanan
d8955d5568 scripts: add description to ceph-release-notes
Signed-off-by: Abhishek Lekshmanan <alekshmanan@suse.com>
2016-10-10 21:39:51 +02:00
Jason Dillaman
3b81f68719 librbd: include pool id in data block prefix if data pool enabled
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 14:08:09 -04:00
Jason Dillaman
cb3712e08c librbd: initialize the data pool IO context during image open state machine
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 14:08:09 -04:00
Jason Dillaman
f5a244ae3d librbd: store data pool id with image header during creation
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 14:08:09 -04:00
Mykola Golub
e4a3d56e34 Merge pull request #11355 from dillaman/wip-17424
rbd: add support for separate image data pool

Reviewed-by: Mykola Golub <mgolub@mirantis.com>
2016-10-10 21:07:44 +03:00
Casey Bodley
4ededdba53 Merge pull request #10731 from yehudasa/wip-rgw-sync-plugins
rgw: sync modules, metadata search

Reviewed-by: Casey Bodley <cbodley@redhat.com>
2016-10-10 13:51:10 -04:00
Jason Dillaman
fe3e2eafa0 librbd: ignore notify errors on missing image header
The rename op on v1 images fails since the header no longer exists. In
the general case, the removal of the header object will also fail the
watcher which has its own recovery path.

Fixes: http://tracker.ceph.com/issues/17549
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 11:57:03 -04:00
Jason Dillaman
97fb2e014e Merge pull request #11363 from xiexingguo/xxg-wip-fix-rbd-warn
cls_rbd: silence compiler warnings

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 11:07:11 -04:00
Sage Weil
870c428004 Merge pull request #11371 from stiopaa1/mds_cinode_removeCount
mds/CInode.h: remove unneeded use of count

Reviewed-by: Sage Weil <sage@redhat.com>
2016-10-10 10:03:25 -05:00
Sage Weil
64daf4cfe6 Merge pull request #11383 from stiopaa1/mds_fsmap_passSharedPtrByConstRef
mds/FSMap: pass shared_ptr by const ref
2016-10-10 10:00:05 -05:00
Jason Dillaman
d54c44350b rbd: include data pool name within info command
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 10:41:57 -04:00
Jason Dillaman
f0d36b7a39 rbd: support overriding image data pool when creating images
Fixes: http://tracker.ceph.com/issues/17424
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 10:41:57 -04:00
Jason Dillaman
d776165f6a Merge pull request #11290 from trociny/wip-17017
rbd-mirror HA: move librbd::image_watcher::Notifier to librbd::object_watcher

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 10:41:25 -04:00
Jason Dillaman
297f21d88b Merge pull request #11260 from runsisi/wip-fix-mirror-image-disable
librbd: fix rollback if failed to disable mirroring for image

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 10:40:33 -04:00
Sage Weil
5f0a360b52 Merge pull request #11301 from liewegas/wip-bluestore-buffered-write
os/bluestore: allow default to buffered write

Reviewed-by: Mark Nelson <mnelson@redhat.com>
2016-10-10 09:40:06 -05:00
Jason Dillaman
99cd30914f Merge pull request #11370 from ceph/wip-krbd-unmap-options
rbd: expose rbd unmap options

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
2016-10-10 10:17:09 -04:00
Sage Weil
3cc40d769b Merge pull request #11255 from dzafman/wip-scrub-boundary
osd: fix scrub boundary to not include a SnapSet

Reviewed-by: Sage Weil <sage@redhat.com>
2016-10-10 09:12:16 -05:00
Sage Weil
0dbe8fd398 msg: make loopback Connection feature accurate all the time
In 626360aab0 we made the
OSD cluster loopback connection CEPH_FEATURES_ALL, but
all other loopback connections got features == 0.  I
can't come up with any reason we wouldn't want those
connections to have accurate feature bits, so let's just
use CEPH_FEATURES_ALL for all of them.

While we're here, make the cflags argument required.

Signed-off-by: Sage Weil <sage@redhat.com>
2016-10-10 09:55:54 -04:00
Sage Weil
e416ccafc1 Merge pull request #11377 from xiexingguo/xxg-wip-bluestore-1008
os/bluestore: drop inline_dirty from struct ExtentMap

Reviewed-by: Sage Weil <sage@redhat.com>
2016-10-10 08:48:43 -05:00
xie xingguo
096811cf06 os/bluestore: narrow scope of 'pos'
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
2016-10-10 19:39:18 +08:00
John Spray
150d383272 Merge pull request #11281 from jcsp/wip-17466
mds: handle bad standby_for_fscids in fsmap

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
2016-10-10 13:21:07 +02:00
xie xingguo
262f2c429e os/bluestore: get rid of multiple call to dirty_blob()
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
2016-10-10 18:52:23 +08:00
Kefu Chai
7f21c6f7fc Merge pull request #10722 from stiopaa1/mon_moncap_addMoveToStrings
mon/MonCap.h: add std::move for std::string

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>
2016-10-10 14:47:21 +08:00