RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-02-12 13:29:17 +00:00
Code Issues Packages Projects Releases Wiki Activity
58,735 Commits 211 Branches 395 Tags 1.9 GiB
700e9c406a
Commit Graph

58735 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Weibing Zhang
bf21c5a7b8 rgw: rgw_rados.cc fix shard_num format for snprintf 
Signed-off-by: Weibing Zhang <zhangweibing@unitedstack.com>
 2016-10-14 16:17:42 +08:00
John Spray
1f3d96c501 test: temporarily disable fork()'ing tests 
Fixes: http://tracker.ceph.com/issues/16556
Signed-off-by: John Spray <john.spray@redhat.com>
 2016-10-14 08:57:47 +01:00
xie xingguo
0ccde6eca3 os/filestore: fix potential result code overwriting 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-14 15:13:02 +08:00
Sage Weil
c701628da2 Merge pull request #11484 from ceph/wip-move-ranges-tp 
tracing/objectstore.tp: add missing move_ranges_... tp
 2016-10-13 19:22:20 -05:00
Sage Weil
9983cd36fa tracing/objectstore.tp: add missing move_ranges_... tp 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-13 16:57:58 -04:00
Patrick Donnelly
20d63508b5
cephfs: add gid to asok status 
This is useful e.g. for building a database with a unique ID for each
entity.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
 2016-10-13 15:59:40 -04:00
Sage Weil
678be45532 os/bluestore: simplify blob print if !compressed 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-13 15:46:17 -04:00
Sage Weil
3ad8ccc53a Merge pull request #11468 from xiexingguo/xxg-wip-bluestore-1013 
os/bluestore: traverse buffer_map in reverse order when spliting BufferSpace

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-10-13 14:21:10 -05:00
Sage Weil
62c12b70e2 Merge pull request #11296 from liewegas/wip-fixes 
osd: drop stray debug message
 2016-10-13 14:00:45 -05:00
Sage Weil
907fde0c26 Merge pull request #11237 from liewegas/wip-ec-move 
os: move_ranges_destroy_src

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-10-13 13:23:48 -05:00
Sage Weil
6244755a70 mon/OSDMonitor: set REQUIRE_JEWEL when setting REQUIRE_KRAKEN 
Lots of checks look for the jewel flag; setting the
kraken flag should also set this one.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-13 12:28:18 -04:00
Sage Weil
12e508313d mon/OSDMonitor: health warn if require_{jewel,kraken} flags aren't set 
We want to prompt users to set these flags as soon as their
upgrades complete.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-13 12:16:40 -04:00
Casey Bodley
32673ba97b Merge pull request #11426 from ceph/wip-rgw-17371 
rgw: loses realm/period/zonegroup/zone data: period overwritten if somewhere in the cluster is still running Hammer

Reviewed-by: Casey Bodley <cbodley@redhat.com>
 2016-10-13 12:05:41 -04:00
Kefu Chai
0f85b344b3 Merge pull request #11391 from agraf/fix-aarch64-crc-v3 
AArch64: Detect crc32 extension support from assembler

Reviewed-by: Yazen Ghannam <yazen.ghannam@linaro.org>
Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-10-13 22:44:04 +08:00
Sage Weil
37c27f044d Merge pull request #11451 from liewegas/wip-bluestore-reshard-fixes 
os/bluestore: fix writes that span existing shard boundaries

Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2016-10-13 09:05:25 -05:00
Sage Weil
c908cae176 os/bluestore: force reshard if new extent spans shard boundary 
A new write may create a new lextent that spans an
existing shard boundary.  If that happens, set a flag
so that we force a reshard when the onode is written.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-13 09:42:17 -04:00
Igor Fedotov
3d3c61d4e8 Merge pull request #11452 from liewegas/wip-bluestore-keys 
os/bluestore: fix onode vs extent key suffix
Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>
Reviewed-by: Igor Fedotov <ifedotov@mirantis.com>
 2016-10-13 15:07:46 +03:00
xie xingguo
4262859b50 os/bluestore: add checking to catch unspecified compression algorithm 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-13 20:00:51 +08:00
xie xingguo
98677bdfe8 os/bluestore: traverse buffer_map in reverse order 
Which is more efficient.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-13 20:00:51 +08:00
Kefu Chai
cd2e18ac30 Merge pull request #11454 from zealoussnow/wip-1013 
doc: fix dead link of "os-recommendations" in troubleshooting-osd

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-10-13 15:31:41 +08:00
xie xingguo
0721e84dc9 common/config_opts: comment nit 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-13 15:04:53 +08:00
xie xingguo
f538a7b91f os/bluestore: kill bluestore_shared_blob_hash_table_size_ratio 
Which is depreciated by 7260166da2 (diff-6519c151b74b5a9355a697be278a4382)

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-13 15:04:53 +08:00
xie xingguo
c701dd61f4 os/bluestore: fix tips when adding buffer to cache 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-13 11:19:18 +08:00
Leo Zhang
f281f8d30e doc: fix dead link of "os-recommendations" in troubleshooting-osd 
Signed-off-by: Leo Zhang <nguzcf@gmail.com>
 2016-10-13 10:38:10 +08:00
Sage Weil
2e11cdb7fc Merge pull request #11439 from xiexingguo/xxg-wip-bluestore-1012 
os/bluestore: avoid polluting shard info if need resharding

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-10-12 17:28:50 -05:00
Sage Weil
d51debaa84 os/bluestore: fix onode vs extent key suffix 
I set the extent key suffix as 'x', thinking that was
not a valid hex character and would let us quickly
identify extent keys (vs onode keys, which end with an
encoded u64). But that doesn't work: the keys encode
integer values in raw form--not in hex
(pretty_binary_string just prints it out that way in
the debug log).

Fix by appending 'o' to the onode keys, so that we will
always have a trailing 'o' or 'x' and can use the last
char reliably to determine the type.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-12 18:27:14 -04:00
Casey Bodley
90b8196b4b Merge pull request #11269 from yehudasa/wip-multipart-uploads-cp 
rgw: multipart upload copy

Reviewed-by: Casey Bodley <cbodley@redhat.com>
 2016-10-12 14:48:28 -04:00
Kefu Chai
09ac5db016 Merge pull request #11403 from bassamtabbara/wip-cmake-allocator-fix 
cmake: find gperftools package for tcmalloc_minimal too

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-10-13 01:01:39 +08:00
Kefu Chai
97aca85fc1 Merge pull request #11442 from wjwithagen/wip-wjw-freebsd-readme 
README.FreeBSD: update to match the bimonthly FreeBSD status report

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-10-13 00:41:02 +08:00
David Zafman
3439cb0a48 Merge pull request #11449 from dzafman/wip-scrub-prep 
test/osd-scrub-repair.sh: Use test case specific object names to help…

Reviewed-by: Kefu Chai <kchai@redhat.com>
 2016-10-12 09:35:40 -07:00
Willem Jan Withagen
7cd32bc328 README.FreeBSD: update to match the bimonthly FreeBSD status report 
Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
 2016-10-12 18:07:43 +02:00
David Zafman
0bf4da5891 test/osd-scrub-repair.sh: Use test case specific object names to help with diagnostics 
Signed-off-by: David Zafman <dzafman@redhat.com>
 2016-10-12 08:36:59 -07:00
Igor Fedotov
b9953a1e20 compressor/ZLibCompressor: fix broken isal-l 
Signed-off-by: Igor Fedotov <ifedotov@mirantis.com>
 2016-10-12 13:38:26 +00:00
Jeff Layton
db2e7e0811 client: lseek shouldn't do permission checking 
With the addition of the UserPerm changes, ceph_lseek started
checking for open permissions using the default mount perms. lseek
should not require those permissions though, as it operates on an
already-opened file descriptor. Remove the perms argument from most
of these functions.

Fixes: be9e43e2da
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
b464161d93 client: convert ceph_setattr callers to ceph_setattrx 
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
09adfcaf54 pybind: fix the libcephfs bindings to use ceph_statx and ceph_fstatx 
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
a283bd2c51 jni: convert libcephfs java bindings to use ceph_statx based APIs 
The java calls themselves are still the same for now. Eventually, it
would be good to covert the java API over to something more statx
like.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
7d78e41b6b client: convert some ceph_fstat callers to ceph_fstatx 
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
209e291813 client: convert some ceph_stat and ceph_lstat calls to ceph_statx 
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:42 -04:00
Jeff Layton
39f4f159fd client: fix caps handling when calling _ll_getattr 
The _ll_getattr was hardcoded to request CEPH_STAT_CAP_INODE_ALL, even
when the caller didn't need that many caps. Have it take a caps mask
as a parameter and pass that through to _getattr.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:41 -04:00
Jeff Layton
bb053b5be4 client: fix handling of ctime and change_attr in fill_statx 
Technically, we need all the shared caps to fill those fields out, to
force synchronization from all the clients. Ensure that we have the
right caps before filling out those fields.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:41 -04:00
Jeff Layton
94e310e945 client: small cleanup in fill_statx() 
Make the setting of stx_ino more compact, and add comment explaining
why we set the type bits there.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:41 -04:00
Jeff Layton
08716fb065 client: fix AT_SYMLINK_NOFOLLOW handling 
We're passing the reverse to path_walk. It should be true if we want
to follow links.

Fixes: f7c885e1f9
Fixes: bce221c1b9
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:41 -04:00
Jeff Layton
7ea6d86fa8 client: remove vestigial uid and gid params from path_walk 
Signed-off-by: Jeff Layton <jlayton@redhat.com>
 2016-10-12 07:04:41 -04:00
Willem Jan Withagen
ea708e749b do_freebsd.sh: Set options for debug building. 
And that requires that the options need to be passed on completely
to the cmake program as well. Thus adding """'s around the option
argument with do_cmake.sh as well

Signed-off-by: Willem Jan Withagen <wjw@digiware.nl>
 2016-10-12 11:06:00 +02:00
xie xingguo
209f058590 os/bluestore: avoid polluting shard info if need resharding 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
 2016-10-12 16:12:14 +08:00
Sage Weil
1906ac9bcf os/bluestore: remove dead code from reshard() 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-10-11 18:36:32 -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
Michal Jarzabek
209e7fb522 mds/DamageTable.cc: move shared ptrs 
Signed-off-by: Michal Jarzabek <stiopa@gmail.com>
 2016-10-11 19:54:16 +01:00
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