Commit Graph

9655 Commits

Author SHA1 Message Date
Sage Weil
4abb47a24c qa: drop numeric prefix from workunits 2010-03-30 21:24:45 -07:00
Sage Weil
b206d2a648 osd: don't filter out shutdown msg to self 2010-03-30 13:40:41 -07:00
Sage Weil
2a0ed8301e msgr: init local_pipe peer_{addr,type} 2010-03-30 13:19:30 -07:00
Sage Weil
9b552c753f Merge branch 'mds' into unstable 2010-03-30 12:22:37 -07:00
Sage Weil
47286cb1ba mds: todo rejoin bug 2010-03-30 12:14:03 -07:00
Sage Weil
3ab157faa5 todo: kclient bug 2010-03-30 12:13:33 -07:00
Greg Farnum
a9f6bd3908 rgw: revert accidental commit of get_attr prototype change 2010-03-30 11:48:06 -07:00
Sage Weil
858314367f Merge branch 'msgr' into unstable 2010-03-30 11:45:00 -07:00
Sage Weil
3edc9d67bd msgr: source, not orig_source, in dbg output 2010-03-30 11:40:38 -07:00
Sage Weil
bf62cc01c4 mds: don't use get_orig_source
MDS doesn't do any request forwarding, so there is no need.
2010-03-30 11:40:38 -07:00
Sage Weil
ddf61d067c msgr: NOSRCADDR protocol feature drops src/orig_src in msg header
This lets us drop the src, orig_src ceph_entity_addr's from the message
header, saving about 160 bytes per message.

The feature is optional.  We can still talk to peers who use the old
protocol.
2010-03-30 11:40:38 -07:00
Sage Weil
9a4b7686c2 msgr: put features in connection_state 2010-03-30 11:40:38 -07:00
Sage Weil
cd102fb61c msgr: use connection for src addr 2010-03-30 11:40:38 -07:00
Sage Weil
94137309db msgr: make Message::get_orig_* differ only when explicitly directed to
This paves the way for removal of the orig_src field from the message
header.
2010-03-30 11:40:38 -07:00
Sage Weil
58ce32cc84 mds: fix MDSTableClient ack double journaling
Do not journal ack unless the tid is registered in the LogSegment.  Once
we journal it, we remove it from the LogSegment list, and once it's
journaled, we remove the pending_commit[tid] entry.

This fixes a bug where the mds got two acks, journaled both of them, and
crashed in the completion for the second because pending_commit[tid] was
gone.  The second ack should have been ignored.
2010-03-30 10:30:02 -07:00
Sage Weil
d5574993f4 mds: start file recovery after sending rejoin ack
The rejoin ack intializes replica lock states correctly; we can't send any
lock messages before that.  This fixes both the check max size call (which
sends lock messages taking the wrlock) and the file_recover() call
(which does the same).

Instead, we make two lists, files to recover and those to fix up.  The lock
states for both are set to PRE_SCAN (LOCK on replica).  After the rejoin
acks go out, we either check_inode_max_size or file_recover.

If file_recover someday grows another caller, this may need something a bit
more sophisticated.
2010-03-29 16:26:54 -07:00
Sage Weil
65d3f3fd83 mds: delay check_inode_max_size if frozen 2010-03-29 15:52:51 -07:00
Sage Weil
e6cd1c219f mds: less noisy about snaprealm siblings
This can get _very_ large for the stray dir.
2010-03-29 14:40:48 -07:00
Greg Farnum
dab367f52f rgw: add optional auid parameter to create_bucket.
If set this will set the RADOS pool's auid to the given one, granting
partial data coherency.
2010-03-29 11:49:11 -07:00
Greg Farnum
66c5af2cca rgw: remove dead code. rebuild_policy exists elsewhere; the acl needs no main() 2010-03-29 11:49:11 -07:00
Greg Farnum
7030731bd8 rgw: radosgw_admin will quit gracefully if it can't open the store 2010-03-29 11:49:11 -07:00
Greg Farnum
a5d809215a rgw: set auid if specified at creation 2010-03-29 11:49:11 -07:00
Greg Farnum
83a6efefe2 rgw: add auid field to RGWUserInfo 2010-03-29 11:49:11 -07:00
Greg Farnum
06393de7ac auth: Keyring no longer segfaults if HOME env variable is unset.
Just skip that path, instead. We can't recover what it meant.
2010-03-29 11:49:11 -07:00
Greg Farnum
9dd4f9923d rgw: if you can't init_storage_provider(), quit instead
of segfaulting later.
2010-03-29 11:49:11 -07:00
Greg Farnum
d9439dd065 rgw: say why the store failed to open in admin 2010-03-29 11:49:11 -07:00
Greg Farnum
06d8e1696f rgw: add minor data to schema examination 2010-03-29 11:49:10 -07:00
Sage Weil
aa490a761b mkcephfs: osd=allow * caps for osd, mds 2010-03-29 09:48:28 -07:00
Sage Weil
9d81f9d699 mds: migrate frag/nest scatterlock info on bounding frags during export
This ensures that the auth inode continues to maintain accurate scatterlock
info about open frags.  We include info on export if it is a bounding frag.
On import, we only take it if we are !auth.  This mirrors the scatterlock
scatter/gather logic in CInode::{encode,decode}_lock_state.
2010-03-26 16:04:45 -07:00
Sage Weil
848caaf05f auth: standardize on -n/--name [type.]name 2010-03-26 16:04:26 -07:00
Sage Weil
c6028ae024 auth: initialized uninitialized variable, don't free pts we don't own 2010-03-26 15:57:17 -07:00
Sage Weil
855bae73d2 mds: make rdlock waiter wait on STABLE | RD, not just RD, when lock is unstable 2010-03-26 14:16:28 -07:00
Sage Weil
a046ca2258 mds: use 32 bits for MDSCacheObject::ref count
16 is not enough!  :)
2010-03-26 13:40:27 -07:00
Sage Weil
cfe60181e9 Merge branch 'unstable' into mds
Conflicts:
	src/TODO
2010-03-26 13:22:04 -07:00
Sage Weil
c5933e4459 mkcephf: behave when user is not defined 2010-03-26 12:34:28 -07:00
Sage Weil
158d4ecb5a auth: fix a few memory leaks, uninit vars
Courtesy of cppcheck
2010-03-26 12:32:01 -07:00
Sage Weil
7975cf0689 mds: skip purge_stray when stray pins are present
Otherwise we can get into a loop:

 eval_stray
 purge_stray
 _purge_stray_purged
  journal truncate
 eval_stray
 ...

This doesn't explain what the stray pin was that I saw, but it at least
avoids the looping.
2010-03-26 12:32:01 -07:00
Sage Weil
3b5d359444 Makefile: add missing MDentryLink.h 2010-03-26 12:32:01 -07:00
Sage Weil
90959633c6 mds: drop unused CInode::PIN_PURGING 2010-03-26 12:32:01 -07:00
Sage Weil
ee127d2330 src/README is required 2010-03-26 12:32:01 -07:00
Sage Weil
20f7f79ab5 filestore: less noisy about OpSequencer queue 2010-03-26 12:32:01 -07:00
Sage Weil
52a908312e mds: drop obsolete hack for base inodes
We used to skip base inodes for scatter_writebehind.  But we can
journal these just like anything else, and it potentially breaks
try_to_expire if a base inode's lock is dirty, because the
completion queued on WAIT_STABLE by scatter_nudge never gets
completed.
2010-03-26 12:32:00 -07:00
Sage Weil
992e429ec2 mds: fix lost completion in scatter_nudge (broke log trimming) 2010-03-26 12:32:00 -07:00
Greg Farnum
21b88da257 rados_bencher: Don't re-initialize passed-in pool, it breaks stuff 2010-03-25 16:18:48 -07:00
Greg Farnum
99e48ee20f osd: Fix up OSDCaps::get_pool_cap to work/make more sense 2010-03-25 16:18:48 -07:00
Greg Farnum
2860eea272 rados: add interface to change pool owners 2010-03-25 16:18:48 -07:00
Greg Farnum
c2e7eed6c7 librados: add change_pool_auid functions 2010-03-25 16:18:48 -07:00
Greg Farnum
1a4899d1cb objecter: add change_pool_auid function.
I'm reluctant to stick this in the objecter since it doesn't quite fit, but
it's a pool management function and putting it here makes it easy to use
elsewhere while maintaining librados' standard function flow.
2010-03-25 16:18:43 -07:00
Sage Weil
a33e9e774f mds: allow rdlock on replica to request sync from auth
This fixes a ping-pong readdir between inode and dirfrag auth.
2010-03-25 13:17:06 -07:00
Greg Farnum
8d9f4fa227 mon: Set MForward::client_caps explicitly; this fixes a bad connection deref.
Previously the decoded PaxosServiceMessage would attempt to follow an
uninitialized Connection* and segfault the system.
2010-03-25 12:58:14 -07:00