Commit Graph

9770 Commits

Author SHA1 Message Date
Sage Weil
96befd4a5a mds: fix broken snaprealm split optimization for (unlinked) files
Splitting a file is trivial; do it.  No special case for 'stray'.
2010-04-14 11:31:52 -07:00
Greg Farnum
6573493a8d rgw: don't crash if HTTP_HOST isn't set 2010-04-13 15:48:01 -07:00
Sage Weil
a6c80a3a1e osd: avoid PGPool lookup in inner loop
Each PG already has a PGPool *.
2010-04-13 12:27:08 -07:00
Sage Weil
69fda2a9c4 osd: fix loading of snapset from _snapdir object attr
This happens when the head is deleted but we still have snaps.  A nested
bufferlist declaration was masking things.
2010-04-13 12:27:08 -07:00
Greg Farnum
e5cd2b72f0 osd: replace the ALLOW_MESSAGES_FROM macro with use of OSDCaps functions.
Now with less null pointer exception! Code checks for session
before dereferencing it.
2010-04-13 11:49:34 -07:00
Greg Farnum
50cad81436 osd: add a message_from_mon function 2010-04-13 11:49:34 -07:00
Greg Farnum
c94e490430 Revert "osd: replace the ALLOW_MESSAGES_FROM macro with use of OSDCaps functions"
This reverts commit 9ceb3be9a6.
This will be coming back in another form shortly, but apparently we have
more connections without sessions on the OSD.
2010-04-13 11:03:45 -07:00
Sage Weil
81e748dffd mds: only recover files from stale clients when auth
Replicas have client_ranges defined too, but only the auth MDS should do
the file size/mtime recovery.

Also rename clear_client_caps_after_export() for clarity, and assert inode
is_auth throughout file recover code.
2010-04-13 10:38:48 -07:00
Greg Farnum
e7547dc7ac osd: add some put()s of Session * that weren't being used properly.
This fixes a memory leak that could have caused problems in the future.
2010-04-12 15:31:16 -07:00
Greg Farnum
9ceb3be9a6 osd: replace the ALLOW_MESSAGES_FROM macro with use of OSDCaps functions 2010-04-12 15:30:13 -07:00
Greg Farnum
4e138c1fc6 osd: Add is_X functions to OSDCaps (where X is mon, osd, mds). 2010-04-12 15:17:15 -07:00
Greg Farnum
17d24513e2 rados: check that given pool name actually exists when deleting 2010-04-12 11:20:12 -07:00
Greg Farnum
d44c26258d mon: OSDMonitor makes sure pool ID is good, preventing crashiness 2010-04-12 11:20:07 -07:00
Sage Weil
c4892bed6f osdmap: move pool names into first part of encoding; add version to second part
This gives the kclient access to pool names while still ignoring the
osd_info_t crap.  And also will make adding more stuff to the second half
later possible without breaking the kclient.
2010-04-09 15:44:50 -07:00
Sage Weil
6bf54f1041 librados: add lookup_pool() 2010-04-09 15:13:37 -07:00
Sage Weil
6417e6a58f librados: no need to include linux/types.h 2010-04-09 14:52:32 -07:00
Sage Weil
26325b1a29 debian: more librados1-dev deps 2010-04-09 14:52:29 -07:00
Sage Weil
414a6e3e87 spinlock: clean up #include deps 2010-04-09 14:52:27 -07:00
Sage Weil
33e604dd6a monclient: seed RNG so we pick a random monitor 2010-04-09 14:52:01 -07:00
Sage Weil
6ed1f04fa2 todo: move bugs to tracker 2010-04-09 14:52:01 -07:00
Greg Farnum
94307c9ab0 auth: cauthtool now identifies itself properly to common_init 2010-04-09 12:35:08 -07:00
Greg Farnum
d5b392474d mon: add 'get' command to AuthMon, letting you retrive one entity's keyring data 2010-04-09 12:35:08 -07:00
Sage Weil
fa9588ab13 filestore: fix weird references for clone/clone_range
Not sure why that was legal to begin with?
2010-04-09 11:50:43 -07:00
Sage Weil
a9ecb46abc objectstore: fix clone_range() transaction encoding 2010-04-09 11:53:28 -07:00
Sage Weil
b4c564c7d9 librados: clean up object listing api 2010-04-08 21:49:17 -07:00
Sage Weil
bb2c0a9298 osd: clear initial interval_set in build_removed_snaps
Caller may pass a non-empty set, in which case we crashed or returned a bad
result.
2010-04-08 21:21:36 -07:00
Sage Weil
1bcdf34816 testrados: create/use own pool 2010-04-08 20:55:36 -07:00
Sage Weil
0f22be7d89 crc32c: ifdef out dead code to shut up warning 2010-04-08 16:42:59 -07:00
Sage Weil
b3cb1cb0d4 Merge branch 'ns' into unstable 2010-04-08 16:42:15 -07:00
Sage Weil
736a271834 filestore: more informative error on current/ rename failure 2010-04-08 16:42:11 -07:00
Sage Weil
c511dd1a42 init-ceph: include 'noatime' btrfs mount option by default 2010-04-08 16:42:11 -07:00
Sage Weil
4a2ca2b194 testrados: test more stuff 2010-04-08 15:52:40 -07:00
Sage Weil
c2987ce40f objecter: cleanup 2010-04-08 15:52:34 -07:00
Sage Weil
f7f38df3d5 librados: clean up c header 2010-04-08 15:52:28 -07:00
Sage Weil
2112d2081d debian: fix up librados1-dev with all necessary headers 2010-04-08 15:23:03 -07:00
Sage Weil
13d6b38c52 librados: move c++ api into librados.hpp; more type cleanup 2010-04-08 15:22:50 -07:00
Sage Weil
1e05d5c099 librados: use std::string, not nstring, in c++ api 2010-04-08 14:19:01 -07:00
Sage Weil
181b81c19a librados: avoid snapid_t type, types.h 2010-04-08 13:00:29 -07:00
Sage Weil
6c9b36b2db rbdtool: fix man warnings 2010-04-08 12:57:46 -07:00
Sage Weil
5c00d5f75a filestore: check for SNAP_DESTROY ioctl, and file if not present 2010-04-08 10:39:47 -07:00
Sage Weil
c9cb118330 librados: move librados.h to include/ 2010-04-08 09:54:18 -07:00
Sage Weil
2fae333d05 spinlock: move to include/ 2010-04-08 09:51:14 -07:00
Sage Weil
ee13f60f97 spinlock: move to ceph namespace 2010-04-08 09:49:50 -07:00
Sage Weil
e1896d8da8 atomic: move to ceph namespace 2010-04-08 09:48:08 -07:00
Sage Weil
3ea82baa07 atomic: rename test() -> read()
Make the interface slightly less weird.
2010-04-08 09:40:10 -07:00
Sage Weil
aa55570d32 buffer: avoid including ceph assert directly
If assert isn't defined, include the standard C one.  Internal ceph bits
should include assert.h before buffer.h if they want the ceph one.
2010-04-08 09:16:11 -07:00
Sage Weil
df2422f364 assert: use ceph namespace 2010-04-08 09:09:43 -07:00
Sage Weil
74be67713d crc32c: prefix with ceph_ 2010-04-08 09:04:44 -07:00
Sage Weil
ac70c56300 page: move to ceph namespace, defs to common/page.cc 2010-04-08 08:53:03 -07:00
Sage Weil
fc8e3c354b buffer: move to ceph namespace 2010-04-08 08:49:27 -07:00