RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-19 01:46:00 +00:00
Code Issues Packages Projects Releases Wiki Activity
41,397 Commits 211 Branches 394 Tags 2.1 GiB
6c2182deb2
Commit Graph

41397 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Sage Weil
7762f1886c auth: optimize crypto++ key context 
The cbc appears to be stateful, but the key is not.

Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
973cd1c00a auth/Crypto: optimize libnss key 
Allocate the key and slot once per key instead of once per encrypt/
decrypt operation.  This improves things by a factor of almost 4x
for my microbenchmark.

Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
16b3515af3 auth: refactor crypto key context 
Cache all of the crypto key context in a new CryptoKeyHandler struct that
is attached to CryptoKey.

Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
fb4b6c52d8 unittest_crypto: benchmark 100,000 CryptoKey::encrypt() calls 
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
8d16d4ce14 auth/cephx: optimize signature check 
The encode_encrypt() helper will generate a bufferlist with a simple
structure.  Profiles seem to indicate this is taking a large amount of time
in the message receive path.

Avoid the encode overhead since we have a tiny buffer with a fixed and
known structure.

Reported-by: Andreas Bluemle <andreas.bluemle@itxperts.de>
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
4e14a5fedb auth/cephx: move signature calc into helper 
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
e972a69511 auth/Crypto: avoid memcpy on libnss crypto operation 
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
e874a9b7d2 do_autogen.sh: unambiguously use nss unless user asks for cryptopp 
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
ad5a1549f4 auth: make CryptoHandler implementations totally private 
There is no need to expose these implementations outside Crypto.cc.

Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 17:00:27 -07:00
Sage Weil
6063a21f9e logrotate.conf: prefer service over invoke-rc.d 
On trusty 14.04, service works but invoke-rc.d does not (but is present).

Fixes: #11330
Reported-by: Wim Heirman <wim@heirman.net>
Signed-off-by: Sage Weil <sage@redhat.com>
 2015-04-21 16:08:09 -07:00
Loic Dachary
8aea730f69 tests: osd-bench.sh must be verbose on failure 
When a command fail as it should but it does not contain the expected
output, display the entire output to confirm it's a false negative.

http://tracker.ceph.com/issues/11441 Fixes: #11441

Signed-off-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 23:35:28 +02:00
Loic Dachary
14cb7b8f59 tests: convert osd-bench.sh to ceph-helpers.sh 
Signed-off-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 23:35:28 +02:00
Loic Dachary
5871781b10 ceph-helpers: implement test_expect_failure 
To display the output in case the command did not fail with the expected
output.

Signed-off-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 23:35:28 +02:00
Sage Weil
67990c57a6 Merge pull request #4399 from trociny/fix-osd_df_avgutil 
mon: osd df: fix average_util calculation

Reviewed-by: Sage Weil <sage@redhat.com>
 2015-04-21 14:27:48 -07:00
Gregory Farnum
1c8913cf2c Merge pull request #4386 from ceph/wip-10657 
mds: warn when clients are not advancing their oldest_client_tid

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
 2015-04-21 14:19:21 -07:00
Gregory Farnum
881cb27426 Merge pull request #4331 from ceph/wip-mds-misc 
Wip mds misc

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
 2015-04-21 14:13:09 -07:00
Josh Durgin
eff0598a85 Merge pull request #4424 from ceph/wip-doc-nuke-setmap 
doc: don't mention ceph osd setmap

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2015-04-21 12:58:54 -07:00
Yehuda Sadeh
7a432f7bdf civetweb: update max num of threads 
Fixes: #10243
cherry-picked upstream fixed into submodule

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2015-04-21 11:08:45 -07:00
Loic Dachary
9dbe55ee43 Merge pull request #4420 from tchaikov/wip-man-fix-desc-in-name 
man: fix the description in NAME section

Reviewed-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 19:34:45 +02:00
Josh Durgin
cbc941cb9b Merge pull request #4409 from ceph/wip-11437 
rgw: don't use end_marker for namespaced object listing

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2015-04-21 10:34:15 -07:00
Josh Durgin
d87fb08220 Merge pull request #4408 from ceph/wip-11436 
rgw: adjust return code if can't find upload

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2015-04-21 10:31:22 -07:00
Josh Durgin
36ff465020 Merge pull request #4407 from ceph/wip-11435 
rgw: fail if parts not specified on complete-multipart-upload

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2015-04-21 10:29:08 -07:00
Ilya Dryomov
3bfdf54ca3 doc: don't mention ceph osd setmap 
ceph osd setmap command was disabled and subsequently removed 5 years
ago.  See 9aadd41b20 ("mon: disable 'osd setmap'") and 6d292397aa
("mon: remove dead 'setmap' code").

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
 2015-04-21 20:27:59 +03:00
Josh Durgin
0c278d7a3c Merge pull request #4423 from ceph/wip-11370 
librbd: better handling for duplicate flatten requests

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2015-04-21 10:23:44 -07:00
Jason Dillaman
ef7e210c3f librbd: better handling for duplicate flatten requests 
A proxied flatten request could be replayed, resulting in a
-EINVAL error code being generated on the second attempt. Filter
out that error if it is known the parent did exist before the
op started.

Fixes: #11370
Backport: hammer
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2015-04-21 13:02:29 -04:00
Sage Weil
86ea4cfbd4 Merge pull request #4372 from xiaoxichen/fix_kv_compact_on_mount 
Fix kv compact on mount

Reviewed-by: Sage Weil <sage@redhat.com>
 2015-04-21 09:31:15 -07:00
Yehuda Sadeh
bfb112406b Merge pull request #4418 from oritwas/wip-11438 
rgw: force content-type header for swift account responses without body

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
 2015-04-21 08:44:52 -07:00
Yan, Zheng
12f1b30947 client: add failure injection for not advancing oldest_tid 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 23:14:22 +08:00
Yan, Zheng
901fd1a5c6 mds: include size of completed_requests in session dump 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 23:14:21 +08:00
Yan, Zheng
dcd9302227 mds: warn when clients are not advancing their oldest_client_tid 
Fixes: #10657
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 23:14:15 +08:00
Orit Wasserman
009664e560 rgw: force content-type header for swift account responses without body 
This regression was introduced in commit 4cfeca03ad

Fixes: #11438

Signed-off-by: Orit Wasserman <owasserm@redhat.com>
 2015-04-21 16:31:57 +02:00
Kefu Chai
b2b443cd4a man: fix the description in NAME section 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2015-04-21 21:12:45 +08:00
Loic Dachary
008b151958 Merge pull request #4370 from guangyy/wip-ec-test-fix 
qa : misc fixes to stabilize test-erasure-code suite

Reviewed-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 13:50:51 +02:00
Kefu Chai
139a57d22d Merge pull request #4411 from hjwsm1989/wip-monstore-pool 
mon: ceph-monstore-tool must do out_store.close() in "store-copy" command

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2015-04-21 19:36:36 +08:00
Loic Dachary
ba5a88050d Merge pull request #4417 from osynge/wip-LimitNOFILE 
Increase max files open limit for OSD daemon.

Reviewed-by: Loic Dachary <ldachary@redhat.com>
 2015-04-21 11:50:19 +02:00
huangjun
d85e0f8c50 tools: ceph-monstore-tool must do out_store.close() 
this change fixes the "store-copy" command.
Like the bug reported in http://tracker.ceph.com/issues/10093.

Signed-off-by: huangjun <hjwsm1989@gmail.com>
 2015-04-21 17:50:03 +08:00
Owen Synge
ebda4ba1c6 Increase max files open limit for OSD daemon. 
Under heavy load the number of file descriptors opened
by the OSD can go beyond the 64K file limit. This patch
increases the default to 128K.

Signed-off-by: Owen Synge <osynge@suse.com>
 2015-04-21 11:31:24 +02:00
Yan, Zheng
a5c067427d client: inlcude inline data in FLUSHSNAP cap message 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:17 +08:00
Yan, Zheng
2fad86e2f1 client: flush dirty snap data before allowing new writes 
this guarantees that new writes do not overwrite existing dirty
snap data.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:17 +08:00
Yan, Zheng
d08d834d53 mds: properly update capability's client_follow 
If snap inode exists, capability's client_follow has already been
updated by MDCache::cow_inode(). So we only update client_follow
when the inode was not COWed.

Besides, we need to check return of SnapRealm::get_snap_following().
It can return CEPH_NOSNAP, which is clearly invalid for client_follows.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:17 +08:00
Yan, Zheng
c9fe13fddc mds: fix Locker::_do_null_snapflush 
the snapid check in Locker::_do_null_snapflush() prevents us from
flushing old snap inodes. The check does not make sense.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:17 +08:00
Yan, Zheng
fa34e04587 mds: do null snap flush when processing embeded cap releases 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:16 +08:00
Yan, Zheng
ab44e5971f mds: make sure snap inode's last match existing snapshots 
MDCache::pick_inode_snap() and snap caps flushing code require
this.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:16 +08:00
Yan, Zheng
fc758ce7f1 mds: remove stray dentry from delayed list when it's queued for purging 
this avoids assert(!dn->state_test(CDentry::STATE_PURGING)) in
eval_stray() being triggered when processing the delayed list.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:16 +08:00
Yan, Zheng
5c09be262c mds: move snap inode to new dirfrag's dirty_rstat_inodes 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:56:16 +08:00
Yan, Zheng
d900252729 mds: keep auth pin on CDir when marking it dirty 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2015-04-21 13:55:43 +08:00
Jianpeng Ma
2e90fa3d5d osd/PG: check scrub state when handle CEPH_OSD_OP_SCRUB_MAP. 
There is a chance which cause scrub stop before replica-node send
scrub-map.

Signed-off-by: Jianpeng Ma <jianpeng.ma@intel.com>
 2015-04-21 11:48:41 +08:00
Yehuda Sadeh
d9bf04fdb6 Merge pull request #4257 from rzarzynski/wip-10683 
rgw: rectify support for GET on Swift's account with "limit" parameter == 0

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
 2015-04-20 17:24:11 -07:00
Yehuda Sadeh
d2a67282e0 rgw: don't use end_marker for namespaced object listing 
Fixes: #11437

Don't deal with end_marker if there's a namespace set for the object
listing. The cur_end_marker.empty() check is incorrect if no end_marker
is provided, but a namespace is set. Just avoid this issue entirely, as
it's not needed.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2015-04-20 16:03:44 -07:00
Yehuda Sadeh
58a144de56 rgw: adjust return code if can't find upload 
Fixes: #11436

Return NoSuchUpload (instead of NoSuchKey).

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2015-04-20 15:01:34 -07:00