Sage Weil
105b6698ac
todo
2009-12-11 12:03:29 -08:00
Sage Weil
6088777609
librados: less noisy
2009-12-11 12:03:24 -08:00
Sage Weil
7c3dc8926d
vstart: fix small bash thing
2009-12-11 08:45:13 -08:00
Sage Weil
072be5caa6
osd: drop backlog on recovery completion (when we mark PG_STATE_CLEAN)
2009-12-11 08:37:19 -08:00
Sage Weil
6258b4fdc6
osd: split pg log too
2009-12-11 08:35:12 -08:00
Sage Weil
894d25291a
osd: make osd thread pool sizes adjustable
...
Rename osd_maxthreads to osd_op_threads to be consistent.
2009-12-11 08:34:27 -08:00
Sage Weil
6ebd8160e7
init-ceph: use pkill, not killall, so that c$type.$date is also killed
2009-12-11 07:53:25 -08:00
Sage Weil
ebc3136587
mon: fix mon injectargs
...
If !connection, we know it's local delivery.
2009-12-11 07:53:25 -08:00
Sage Weil
0deb52013e
mon: fix null pointer ref on connection
2009-12-11 07:53:25 -08:00
Sage Weil
3955dd1c93
crush: fix grammer to parse hash
2009-12-11 07:53:25 -08:00
Greg Farnum
298a2d8280
msgr: lockdep found some not-actually-possible bad dependencies; they've
...
been removed to make it happy
2009-12-10 15:08:50 -08:00
Greg Farnum
243325079b
msgr: Add at least a little documentation
2009-12-10 14:56:44 -08:00
Greg Farnum
82af707d74
msgr: Remove noisy debug output.
2009-12-10 14:33:32 -08:00
Greg Farnum
15f05d407a
msgr: discard_queue now removes empty pipe queues from endpoint's map
...
This prevents bad pointer dereferences
2009-12-10 14:33:32 -08:00
Greg Farnum
eb371aa65e
msgr: Cover qlen with a Spinlock for faster access and to limit blocking
2009-12-10 14:33:32 -08:00
Greg Farnum
ec2720cc9e
msgr: That unlock's not in the right spot
2009-12-10 14:33:32 -08:00
Greg Farnum
2e78035ef4
msgr: Fix a number of locking and deletion bugs. Appears to work.
2009-12-10 14:33:32 -08:00
Greg Farnum
3a7ad1b89d
msgr: Fix blocking locks and another potential block.
2009-12-10 14:33:31 -08:00
Greg Farnum
c58f0e90e2
msgr: rename locks for clarity; move some around to prevent blocking.
2009-12-10 14:33:31 -08:00
Greg Farnum
0792302a1d
msgr: implement round-robin pipe message delivery.
2009-12-10 14:33:31 -08:00
Greg Farnum
ec0031b763
msgr: Remove Endpoint::dispatch_queue and queue_message;
...
set up local pipe and necessary routing code.
Also, fix potential locking issue by pushing message onto
queue before pushing queue to round-robin
2009-12-10 14:33:31 -08:00
Greg Farnum
c01741421d
msgr: Add queues, etc needed for message delivery by round-robin pipe
2009-12-10 14:33:31 -08:00
Greg Farnum
eb1be2babe
msgr: rename pipe's outbound message queue
2009-12-10 14:33:31 -08:00
Sage Weil
2756e436e0
config: parse 'lockdep' option in .conf
2009-12-10 13:01:35 -08:00
Sage Weil
429ac4ef9c
ceph.conf.sepia: more osds, lockdep
2009-12-10 13:01:27 -08:00
Sage Weil
1eabf8d188
mon: log error, don't abort, on stray pg_stat notify
2009-12-10 10:03:44 -08:00
Sage Weil
0b83b541c4
osd: set pg_stat.created
2009-12-10 09:56:54 -08:00
Sage Weil
4efe6aeea6
osd: respond to query on uncreated pg with empty info; do not create
...
We should never create a PG* for an uncreated pg except through the mkpg
path.
2009-12-10 09:46:15 -08:00
Sage Weil
cfda3be56f
mon: stop trying to create localized pgs >= max osd|device
...
If lpg_num == lpgp_num, and the preferred osd no longer exists, there is
no guarantee that the parent and child pgs will land on the same osd. This
happens when we are creating pgs for a new osd, they don't finish getting
created, and then we destroy that OSD again.
Care should be taken to avoid expanding the OSD count _and_ lpg_num
simultaneously...
2009-12-10 09:24:31 -08:00
Sage Weil
6743d3f604
todo
2009-12-09 16:06:58 -08:00
Sage Weil
60373a1151
osd: fix mark_down on heartbeat_messenger addrs
...
This fixes residual outgoing connect attempts and
subsequent 'wrong node' messages from old outgoing
osd_ping messages that never leave our queues.
2009-12-09 16:06:57 -08:00
Sage Weil
888236906a
filestore: avoid string copy after collection_list sort
2009-12-09 13:38:25 -08:00
Sage Weil
0bc006e7a2
osd: don't forcefeed bad (large) device ids to crush
2009-12-09 11:58:55 -08:00
Sage Weil
5b6ca82e56
filestore: test xattrs on basedir/fsid, not basedir
...
The basedir might have differing permissions.
2009-12-09 10:43:37 -08:00
Sage Weil
4c1f64acb2
ceph.conf.cosd changes
2009-12-09 10:27:56 -08:00
Sage Weil
4a98c4615b
osd: drop old osdmap compatibility cruft
...
This is old enough to have been superceded by a full disk format rev.
2009-12-09 10:27:56 -08:00
Sage Weil
823513f88d
osd: fix decoding for old osdmaps (pre osd hb change)
...
Including ugly hack to work around badly encoded v3 maps
2009-12-09 10:27:56 -08:00
Sage Weil
70d22ba169
mds: mark down stray connections on reset
2009-12-08 11:58:31 -08:00
Sage Weil
cff5512a66
vstart: add --smallmds option
2009-12-08 08:25:34 -08:00
Sage Weil
9bc49228b0
logger: avoid dup lines when we get behind
...
If we get behind (say, after SIGSTOP ... SIGCONT),
instead of printing dup lines for every interval we missed,
print just the first and last. That will make any graph
have a nice, hopefully obvious perfectly horizontal line.
2009-12-08 08:25:21 -08:00
Yehuda Sadeh
9922ae1f55
qa: add bonnie++ to the workunits
2009-12-07 16:48:35 -08:00
Sage Weil
616a655c22
filestore: limit size of flusher queue
...
Limit the size of the flusher queue to avoid hitting max open files
limit. If we hit the max, just don't queue async flush, so we close the
fd immediately.
2009-12-07 15:43:31 -08:00
Sage Weil
222c6c2027
osd: rev cluster protocol to match osd hearbeat change
2009-12-07 13:47:50 -08:00
Greg Farnum
e0862fd0e0
msgr: Only allow one local endpoint; now has simpler, less iterative code!
2009-12-04 18:23:12 -08:00
Greg Farnum
1ebcebf6ff
osd: Now uses separate messenger for heartbeats.
...
Includes changes to OSDMonitor and OSDMap so things keep working
2009-12-04 16:03:12 -08:00
Greg Farnum
31ebbe50a3
rados: benching sequential reads can now have a time limit
2009-12-04 15:54:28 -08:00
Greg Farnum
a0d3b8f885
rados: change benchmark interface and record objects written/size to osds
2009-12-04 15:54:28 -08:00
Sage Weil
ff4e1552dc
mds: debug reconnect timeouts
2009-12-04 12:07:31 -08:00
Sage Weil
3d97239b2d
debian: only upload tarball whem building amd64
...
For some reason lenny32-packager includes binaries in
the 'make dist' tarball. Lame.
2009-12-04 12:07:18 -08:00
Sage Weil
9bea9d0f6b
logger: use proper format for s64
2009-12-04 11:27:07 -08:00
