1
0
mirror of https://github.com/ceph/ceph synced 2025-03-01 14:03:04 +00:00
Commit Graph

15851 Commits

Author SHA1 Message Date
Sage Weil
8ce6544764 osd: fix _activate_committed() crash
Do not dereference acting[0] unless we know it is still valid.

Take a reference when scheduling the transaction, and drop it in the
completion, to ensure that the PG isn't removed out from underneath us.

Fixes: 
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-18 11:16:51 -07:00
Sage Weil
6d90796fc9 compatset: dump json
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-18 10:07:12 -07:00
Sage Weil
d21c918b73 uninline a bunch of buffer.h stuff 2011-08-18 08:44:28 -07:00
Sage Weil
f6ff0f14e9 uninline osd types 2011-08-18 08:44:24 -07:00
Sage Weil
df99cd9b11 mds: add debug_mds_locker
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-08-17 18:38:20 -07:00
Sage Weil
a724bc4667 librbd: rbd_list: return buffer size, not element count
Fixes: 
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 16:54:46 -07:00
Sage Weil
92746916dc mon: add/remove data_pg_pools from mdsmap
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 16:53:48 -07:00
Samuel Just
a88c1790ff PG: last_warm_restart must be set to the current epoch
Previously, last_warm_restart did not get set until the
first time the pg was restarted.

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
2011-08-17 16:35:53 -07:00
Sage Weil
47d75af51b mds: explain weird encoding
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 16:12:14 -07:00
Sage Weil
a21a5ef8d6 mds: whitespace
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:53:22 -07:00
Sage Weil
b4fb82bb46 mds: drop unused projected dirfrag_t arg for EMetaBlob::add_primary_dentry
Not used.  If it is later, we need to project it properly.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:47:55 -07:00
Sage Weil
93b7167221 mds: drop unused snapbl arg
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:43:21 -07:00
Sage Weil
d011c62394 mds: avoid explicit passing of projected_xattrs
No need to pass this in explicitly; we can look in the projected inode for
it.  This actually fixes a race where a journaled inode following a
setxattr will not journal the projected xattrs.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:41:06 -07:00
Sage Weil
a9998c8d96 mds: avoid passing snapbl around explicitly
We now properly track projected snaprealm/sr_t's.  Use that for journaling
instead of manually passing around a snapbl.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:41:06 -07:00
Sage Weil
08fb8a614d mds: use snaprealm projection for MDCache::snaprealm_create()
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:33:28 -07:00
Sage Weil
6ee52e7d42 mds: simplify sr_t journaling
Drop encode/decode in SnapRealm and encode the sr_t's directly.  This way
the caller says what they mean (current or projected).

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:33:11 -07:00
Sage Weil
d60d5319ad mds: journal inode on mksnap
We were cowing but not journaling a simply-dirty inode.  Der.

Fixes: 
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 15:32:02 -07:00
Sage Weil
79db9fc509 mds: handle O_TRUNC when size is already 0
We always want to go through the truncation path.  Two reasons:

 - even if the size is already 0, we still need to update ctime/mtime
 - we may not have the correct size (client could have dirty data), in
   which case we still want to bump truncate_seq etc.

Basically, we can't trust pi->size here because we are holding a wrlock,
not an xlock.  And that's fine, since we need to update the inode
unconditionally anyway.

This broke cfuse pjd open/00.t tests when we added the fuse option
atomic_o_trunc.  libceph has always been broken in this regard.

Fixes: 
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 14:11:15 -07:00
Sage Weil
45740cf86a client: send O_TRUNC open requests to MDS
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 14:00:08 -07:00
Sage Weil
266110d672 client: move Inode operator<<
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 13:59:57 -07:00
Sage Weil
0ed2fafce0 rbd: remove dup 'rm' command from man page
Fixes: 
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 12:56:09 -07:00
Sage Weil
d314506442 cfuse: take out accidental bad channel add
This was a brainstorm that snuck into last commit, grr.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 12:53:06 -07:00
Yehuda Sadeh
d9df58e1b4 rgw: fix virtual bucket name handling 2011-08-17 12:56:19 -07:00
Sage Weil
2df3f2c946 cfuse: make use of fuse invalidate callback optional (and off by default)
This is hitting a deadlock in fuse itself (or so it appears).  Disable it
until that's sorted out.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 12:25:56 -07:00
Yehuda Sadeh
1179a31b74 formatter: remove bad asserts 2011-08-17 12:16:37 -07:00
Yehuda Sadeh
afdb1f3842 rgw: list objects shows time in GMT, thread safety fix 2011-08-17 11:42:18 -07:00
Colin Patrick McCabe
a311715203 src/perfglue/heap_profiler.cc: fix snprintf
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-17 11:08:34 -07:00
Yehuda Sadeh
14de6d2688 client: compile with older libfuse2 2011-08-17 11:21:55 -07:00
Colin Patrick McCabe
82baaeb172 Merge branch 'wip-librados-tests' 2011-08-17 11:00:58 -07:00
Sage Weil
cb1f05009d osd_types: fix merge
Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 10:47:39 -07:00
Sage Weil
46645936c5 Merge commit 'v0.33' 2011-08-17 10:47:02 -07:00
Sage Weil
340b4348bf v0.33 2011-08-17 09:16:27 -07:00
Sage Weil
89eeb34e65 osd: pg_stat_t: throw exception on unknown encoding
This is better then decoding nonsense, eating ram, etc.

Signed-off-by: Sage Weil <sage@newdream.net>
2011-08-17 09:16:27 -07:00
Sage Weil
ce00564049 qa: test rename into snapped dir
Tests fix for .

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-08-17 09:07:08 -07:00
Sage Weil
e98669ea69 mds: mark newly created snaprealms as open
Fixes: 
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
2011-08-17 09:07:06 -07:00
Colin Patrick McCabe
2dd5510c02 test/rados-api/misc.cc: test Rados::version
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:55:19 -07:00
Colin Patrick McCabe
d45b0f6336 test/rados-api/list.cc: C++ version
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:53:48 -07:00
Colin Patrick McCabe
b8b14c9c09 test/rados-api/snapshots.cc: C++ version
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:48:04 -07:00
Colin Patrick McCabe
65626edec7 test/rados-api/watch_notify.cc: implement C++ ver
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:37:38 -07:00
Colin Patrick McCabe
9ca1527b65 test/rados-api/stat.cc: implement C++ tests
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:30:34 -07:00
Colin Patrick McCabe
d8da2ce6e4 rados/librados.h: use rados_snap_t consistently
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 17:16:09 -07:00
Colin Patrick McCabe
80a094d254 test/rados-api/aio.cc: test RoundTripWriteFull
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 13:06:09 -07:00
Colin Patrick McCabe
9df460593a test/rados-api/aio.cc: implement FlushPP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 12:59:40 -07:00
Colin Patrick McCabe
4d9dab04b1 test/rados-api/aio.cc: test ReturnValuePP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 11:25:39 -07:00
Colin Patrick McCabe
8cc794b087 test/rados-api/aio.cc: test IsSafePP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 11:10:53 -07:00
Colin Patrick McCabe
3b81656a59 test/rados-api/aio.cc: add IsCompletePP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 10:49:34 -07:00
Colin Patrick McCabe
f4d7382932 test/rados-api/aio.cc: Add RoundTripAppendPP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 10:40:25 -07:00
Colin Patrick McCabe
cc25266957 test/rados-api/aio.cc: add RoundTripPP
Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-16 10:31:07 -07:00
Colin Patrick McCabe
8072bdcaa4 Fix bug in turning off logging. Remove log_dir.
Fix a bug that made it impossible to turn off logging to a file.

Remove the old "log_dir" setting. It has been deprecated for a long
time, and it adds a lot of complexity to the code.

heap_profiler: use the directory that log_file is in to construct a
profiler output file name. Don't use log_dir.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
2011-08-15 20:58:23 -07:00
Sage Weil
e7de7cba9d Merge branch 'next' 2011-08-15 16:34:04 -07:00