RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-01-19 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity
50,990 Commits 216 Branches 394 Tags 1.8 GiB
740b95051b
Commit Graph

50990 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Sage Weil
b413a8f455 cls_rgw: use sized uint64_t for encoded time_t 
The size of time_t varies between architectures--do not encode it
directly.

Reported-by: Eric Lee <eric.lee@hgst.com>
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-09 09:48:15 -05:00
Signed-off-by: Eric Lee
f42b86b263 test/librgw_file_*: use size_t* for rgw_read arg 
Fixes 32-bit build.

Signed-off-by: Eric Lee <eric.lee@hgst.com>
 2016-03-09 09:47:48 -05:00
Yan, Zheng
e515f1423d client: pass proper old snap context to queue_cap_snap() 
the snap realm for split point can be newly created, its snap context
is empty in this case.

Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2016-03-09 22:06:25 +08:00
Yan, Zheng
bd611ab189 client: flush kernel pagecache before creating snapshot 
Fixes: #10436
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2016-03-09 22:06:12 +08:00
xie xingguo
1ea8769ba7 osd: remove up_thru_pending field, which is never used 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-03-09 19:17:46 +08:00
Loic Dachary
d29f610b9b Merge pull request #7960 from SUSE/wip-docdev-mon 
doc/dev: continue writing Testing in the cloud chapter

Reviewed-by: Loic Dachary <ldachary@redhat.com>
 2016-03-09 16:41:14 +07:00
Nathan Cutler
323c267971 doc/dev: Continue writing Testing in the cloud chapter 
Signed-off-by: Nathan Cutler <ncutler@suse.com>
 2016-03-09 09:23:39 +01:00
Mykola Golub
27f2d806f9 stop.sh: make more portable 
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:51:43 +00:00
Mykola Golub
88c666d1df compat: don't call gcc explicitly 
`$(CC) -dumpversion` works for clang too, as it supports gcc
compatibility for -dumpversion option.

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:51:43 +00:00
Mykola Golub
2be4029212 rbd: build rbd binary on all platforms 
(excluding linux specific features)

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:51:43 +00:00
Mykola Golub
aa338379ae common: build with blkdev on all platforms 
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:51:43 +00:00
Mykola Golub
6ce4744a5a compat: fix build on non-linux 
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:50:43 +00:00
Mykola Golub
a5722264fd common: include malloc.h only on linux 
On freebsd <malloc.h> has been replaced by <stdlib.h>

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:50:43 +00:00
Mykola Golub
48a1ba58af rocksdb: more portable make invocation 
- export CC and CXX, so the same compiler is used
- use NPROC

Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:50:43 +00:00
Mykola Golub
1ae5600307 freebsd: use sysctl instead of nproc 
Signed-off-by: Mykola Golub <mgolub@mirantis.com>
 2016-03-09 07:50:43 +00:00
Kefu Chai
18e33f243e buffer: add symmetry operator==() and operator!=() 
* add const_iterator(const iterator&) so an iterator can be converted to
  const_iterator
* add tests for operator==() and operator!=()

Signed-off-by: Kefu Chai <kchai@redhat.com>
 2016-03-09 15:35:17 +08:00
Yan, Zheng
9e3441733c ceph_test_libcephfs: shutdown without closing file/dir 
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2016-03-09 14:40:06 +08:00
Yan, Zheng
8fc1b1b095 client: close opened dirs when umounting 
Fixes: #14996
Signed-off-by: Yan, Zheng <zyan@redhat.com>
 2016-03-09 14:40:06 +08:00
x11562
f4b210e79e rgw: remove unused vector 
Signed-off-by: Na Xie <xie.na@h3c.com>
 2016-03-09 10:50:41 +08:00
Josh Durgin
2c01f91168 Merge pull request #7986 from liewegas/wip-xxhash 
xxhash: use clone of xxhash.git; add .gitignore

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2016-03-08 15:22:26 -08:00
Yehuda Sadeh
d7e42213b0 rgw: handle error when fetching data log changes 
Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 14:14:05 -08:00
Sage Weil
aab91667c4 xxHash: add .gitignore for build artifacts 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 16:18:00 -05:00
Sage Weil
517a9d6d52 .gitmodules: use github.com/ceph/xxHash clone 
This lets us add our own commits, if necessary.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 16:04:31 -05:00
Sage Weil
bdcc84042c qa/workunits/rest/test.py: add confirmation to 'mds setmap' 
Broken when we added the confirmation in
fc3554e36a

Fixes: #14606
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 14:34:40 -05:00
Sage Weil
433afce628 mon: initialize last_* timestamps on new pgs to creation time 
Currently, when you create a pool, until the PGs peer you generate a health
error like

            8 pgs are stuck inactive for more than 300 seconds

which is inaccurate and misleading.  Instead, set the timestamps to the
creation time so that warnings don't appear until it's clear they're stuck.

Fixes: #14952
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 13:46:58 -05:00
Jason Dillaman
b705a71b9f librbd: refresh image if required before replaying journal ops 
Fixes: #14908

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 13:20:24 -05:00
Yehuda Sadeh
1a800f6095 rgw: rework completion notifier and manager lifecycle 
completion manager is now refcounted, and keeps track of all the
notifiers. This is needed so that when we shut down we can release
all completion notifiers, so that they don't reference the manager
anymore.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:55 -08:00
Yehuda Sadeh
9106e8ad31 rgw: reorder RGWRados::finalize() 
first stop all sync threads, then delete them. Might be referencing
each other.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:47 -08:00
Yehuda Sadeh
ea29b71166 rgw: don't equeue async cr rados operations if going down 
Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:46 -08:00
Yehuda Sadeh
5914ae9d6f rgw: abort early in realm reload if already reloading 
Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:46 -08:00
Yehuda Sadeh
03d60e3920 rgw: silence some debug messasges 
Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:46 -08:00
Yehuda Sadeh
5ecacfe609 rgw: stop async_rados thread first 
Need to shut this thread first, so that async callbacks don't
occur when we shut down the data/meta sync processors.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
 2016-03-08 10:02:46 -08:00
Sage Weil
1c99dc2f4f doc/release-notes: v10.0.4 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:37:12 -05:00
Sage Weil
bc565e054e Merge pull request #7966 from dachary/wip-release-notes 
release-notes: draft v10.0.4 release notes
 2016-03-08 11:33:52 -05:00
Sage Weil
8384f16de2 Merge pull request #7880 from SUSE/wip_keyring_permisions_for_mon_deamon_14950 
mon: fix keyring permissions

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:22:08 -05:00
Sage Weil
46007d271f Merge pull request #7121 from efirs/ef_eversion_t_no_sprintf 
osd: Replace snprintf with faster implementation in eversion_t::get_key_name

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-03-08 11:21:42 -05:00
Sage Weil
34b7c824f9 Merge pull request #7141 from jecluis/wip-mon-clockskew 
mon: make clock skew checks sane

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-03-08 11:21:20 -05:00
Sage Weil
b533cf107a Merge pull request #7899 from liewegas/wip-osd-markdown 
test/osd: Relax the timing intervals in osd-markdown.sh
 2016-03-08 11:20:54 -05:00
Sage Weil
84caba5277 Merge pull request #7900 from liewegas/wip-bluefs-alignment 
osd: bluefs: fix alignment for odd page sizes

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:20:43 -05:00
Sage Weil
bdd2ddb387 Merge pull request #7904 from liewegas/wip-log-delete 
log: use delete[]

Reviewed-by: Kefu Chai <kchai@redhat.com>
Reviewed-by: Nathan Cutler <ncutler@suse.com>
 2016-03-08 11:20:10 -05:00
Sage Weil
7139a232d2 osd: handle boot racing with NOUP set + clear 
Resolve this race:

- osd sends boot message
- mon sets NOUP
- mon receives and drops boot message
- mon later clear NOUP
-> osd still waiting

by restarting the boot process if we are is_booting() and we see
the NOUP flag cleared.  This implies we sent the boot before NOUP
was set (we don't send boot while it is) and the mon probably
dropped our request.  If it didn't we'll go through another boot
cycle, but that is rare, mostly harmless, and unavoidable (there's
no easy way to tell whether our message was dropped or not).

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:19:26 -05:00
Sage Weil
7eba5baedc osd: inline advance_map() 
This has shrunk over time and is short enough it's easier to
follow the code with it inline.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:19:26 -05:00
Jason Dillaman
4ded44aa5e journal: possible race condition during fetch playback 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 11:01:33 -05:00
Jason Dillaman
3982895286 test: fix errors introduced by rebase to master 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 10:35:41 -05:00
Jason Dillaman
b37f135c64 journal: clean up playback notification handling 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 10:08:13 -05:00
Jason Dillaman
b710374001 tests: updated test cases for librbd journal tag allocation 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 09:03:44 -05:00
Jason Dillaman
21d959529c journal: properly handle tag transition 
Now that the tag concept has been re-used for delineating epochs for
librbd, we need playback to properly handle the cases where the active
playback tag abruptly ends and a newer tag is inserted in the first
splay offset object.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 09:03:44 -05:00
Jason Dillaman
3325b871ee journal: additional debug messages surrounding commit updates 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 09:03:43 -05:00
Jason Dillaman
8444da8480 librbd: improved debugging for journal client / tag processing 
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 09:03:43 -05:00
Jason Dillaman
16d5e695c3 librbd: allocate new journal tag when acquiring exclusive lock 
The journal tag tracks a write epoch within the image, so before
allowing writes to proceed, ensure the journal entries will be
stored under a new epoch.

Signed-off-by: Jason Dillaman <dillaman@redhat.com>
 2016-03-08 09:03:43 -05:00