]> git.apps.os.sepia.ceph.com Git - ceph.git/log
ceph.git
15 years agomds: get rid of old lease mask cruft
Sage Weil [Sat, 24 Jul 2010 04:02:47 +0000 (21:02 -0700)]
mds: get rid of old lease mask cruft

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: fix lease revocation mask vs type
Sage Weil [Sat, 24 Jul 2010 03:51:58 +0000 (20:51 -0700)]
mds: fix lease revocation mask vs type

15 years agoinitscripts: fix type.id check in ceph.conf
Sage Weil [Fri, 23 Jul 2010 23:05:20 +0000 (16:05 -0700)]
initscripts: fix type.id check in ceph.conf

15 years agomds: drop remove_client_lease return value
Sage Weil [Fri, 23 Jul 2010 22:41:46 +0000 (15:41 -0700)]
mds: drop remove_client_lease return value

Unused

15 years agomds: fix dentry lease mask/lockid disconnect
Sage Weil [Fri, 23 Jul 2010 22:39:53 +0000 (15:39 -0700)]
mds: fix dentry lease mask/lockid disconnect

There is only one lock here, no need to be tricky with the mask.

15 years agomds: no return value for issue_client_lease()
Sage Weil [Fri, 23 Jul 2010 22:15:40 +0000 (15:15 -0700)]
mds: no return value for issue_client_lease()

Return value is unused.  Also drop obsolete inode lease method declaration.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: fix lease issue mask
Sage Weil [Fri, 23 Jul 2010 22:14:43 +0000 (15:14 -0700)]
mds: fix lease issue mask

We define 1 to be the only lease "mask" we currently support (for dentry)
and divorce ourselves from the CEPH_LOCK namespace for this purpose.  We
did this in c8d7c970e864e9f130b7c44b29c322771067026e, but screwed up the
issue function.  Fix the caller, and pay attention to @mask here.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: create leases with non-zero seq
Sage Weil [Fri, 23 Jul 2010 21:54:44 +0000 (14:54 -0700)]
mds: create leases with non-zero seq

We want to avoid the situation:

- mds issues lease, seq 0
- mds times out lease, seq 0
- client sends seq 0 release
- mds issues new lease, seq 0 (again.. poor choice of seq#)
- mds processes client lease seq 0 release (bad)

15 years agoosd: log/debug hb set changes
Sage Weil [Fri, 23 Jul 2010 21:41:07 +0000 (14:41 -0700)]
osd: log/debug hb set changes

15 years agoosdmap: recalc num_osds after applying incremental
Sage Weil [Fri, 23 Jul 2010 21:40:59 +0000 (14:40 -0700)]
osdmap: recalc num_osds after applying incremental

15 years agoosd: clean_up_local should ignore snapped, snapdir objects
Sage Weil [Fri, 23 Jul 2010 19:59:11 +0000 (12:59 -0700)]
osd: clean_up_local should ignore snapped, snapdir objects

The snap_trimmer is responsible for removing any/all snap objects.

snapdir only exists if snapped objects exist and head doesn't, so it will
get cleaned up when the snapped object(s) do.

15 years agofilestore: make stat print correct errno
Sage Weil [Fri, 23 Jul 2010 18:08:48 +0000 (11:08 -0700)]
filestore: make stat print correct errno

15 years agoosd: simplify heartbeat checks
Sage Weil [Thu, 22 Jul 2010 22:55:23 +0000 (15:55 -0700)]
osd: simplify heartbeat checks

- Only check heartbeats when we have heartbeat_lock and osdmap rdlocked,
and thus _know_ heartbeat info and map are in sync.  Drop unnecessary
consistency checks.
- Check heartbeats in tick(), since the heartbeat thread may miss it if
it's unlucky.

15 years agoosd: red is good enough; don't blink
Sage Weil [Thu, 22 Jul 2010 21:36:00 +0000 (14:36 -0700)]
osd: red is good enough; don't blink

15 years agomds: only try and lssnaps on a dir for as long as it's been around
Greg Farnum [Fri, 23 Jul 2010 20:42:07 +0000 (13:42 -0700)]
mds: only try and lssnaps on a dir for as long as it's been around

15 years agoqa: runoncfuse shouldn't start cfuse in background any more.
Greg Farnum [Tue, 20 Jul 2010 22:53:31 +0000 (15:53 -0700)]
qa: runoncfuse shouldn't start cfuse in background any more.

15 years agomds: fix snap lookups in CInode encoding
Greg Farnum [Tue, 20 Jul 2010 22:52:43 +0000 (15:52 -0700)]
mds: fix snap lookups in CInode encoding

15 years agomon: don't crash on 'osd pool set' command
Sage Weil [Fri, 23 Jul 2010 20:21:36 +0000 (13:21 -0700)]
mon: don't crash on 'osd pool set' command

15 years agoosd: zero ondisklog pointers when starting pg deletion
Sage Weil [Fri, 23 Jul 2010 18:45:26 +0000 (11:45 -0700)]
osd: zero ondisklog pointers when starting pg deletion

This fixes a problem where the osd stops part way through pg cleanup.  It
seens the old ondisklog bounds, but then fails to load the log, and crashes
on startup.

We just need to zero the ondisklog bounds when we zero the log.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agolibrados: remove use of snapid_t
Yehuda Sadeh [Fri, 23 Jul 2010 18:43:25 +0000 (11:43 -0700)]
librados: remove use of snapid_t

15 years agoobjecter: default debug level 0 (no 'laggy' messages)
Sage Weil [Fri, 23 Jul 2010 16:42:51 +0000 (09:42 -0700)]
objecter: default debug level 0 (no 'laggy' messages)

15 years agoinitscripts: use $type.$id, not $name, when possible
Sage Weil [Thu, 22 Jul 2010 21:27:08 +0000 (14:27 -0700)]
initscripts: use $type.$id, not $name, when possible

15 years agoinitscripts: accept osd0 or osd.0, regardless of what conf uses
Sage Weil [Thu, 22 Jul 2010 21:26:49 +0000 (14:26 -0700)]
initscripts: accept osd0 or osd.0, regardless of what conf uses

15 years agomds: be careful obeying REQRDLOCK
Sage Weil [Thu, 22 Jul 2010 19:01:11 +0000 (12:01 -0700)]
mds: be careful obeying REQRDLOCK

Only do a simple_sync() if we are stable, auth, and not already sync.  The
client request can race with other state changes, so be careful.  The
client will also retry on any state change, so we can safely ignore if
things don't look quite right.

This fxes crash:
mds/Locker.cc: In function 'bool Locker::simple_sync(SimpleLock*, bool*)':
mds/Locker.cc:2592: FAILED assert(0)
1: (Locker::handle_file_lock(ScatterLock*, MLock*)+0x2a8) [0x57d328]
2: (MDS::_dispatch(Message*)+0x2165) [0x4a11d5]
3: (MDS::ms_dispatch(Message*)+0x6d) [0x4a169d]
4: (SimpleMessenger::dispatch_entry()+0x733) [0x47c903]
5: (SimpleMessenger::DispatchThread::entry()+0x1c) [0x474bac]
6: (Thread::_entry_func(void*)+0xa) [0x48748a]
7: /lib/libpthread.so.0 [0x7f8af92ff73a]
8: (clone()+0x6d) [0x7f8af852569d]

15 years agodebian: fix typo in libcrush description
Sage Weil [Thu, 22 Jul 2010 17:53:14 +0000 (10:53 -0700)]
debian: fix typo in libcrush description

Reported-by: Wido
15 years agomds: initialize snaprealm created, current_parent_since on creation
Sage Weil [Tue, 20 Jul 2010 22:59:11 +0000 (15:59 -0700)]
mds: initialize snaprealm created, current_parent_since on creation

Need to initialize created and current_parent_since on new snaprealms
when they are created, or else we get incorrect results from the likes of
SnapRealm::get_snap_info() (e.g., parent snaps from before we were
created).

15 years agoMerge remote branch 'origin/debian' into unstable
Sage Weil [Wed, 21 Jul 2010 20:30:02 +0000 (13:30 -0700)]
Merge remote branch 'origin/debian' into unstable

15 years agorgw: fix calls to finish_get_obj() with uninitialized handle
Yehuda Sadeh [Wed, 21 Jul 2010 18:03:03 +0000 (11:03 -0700)]
rgw: fix calls to finish_get_obj() with uninitialized handle

15 years agodebian: drop unused *.modules.in
Sage Weil [Wed, 21 Jul 2010 16:31:57 +0000 (09:31 -0700)]
debian: drop unused *.modules.in

15 years agodebian: no more ceph-kclient-soruce
Sage Weil [Wed, 21 Jul 2010 04:40:34 +0000 (21:40 -0700)]
debian: no more ceph-kclient-soruce

15 years agodebian: cleanup from Clint Byrum <clint@ubuntu.com>
Sage Weil [Wed, 21 Jul 2010 04:40:02 +0000 (21:40 -0700)]
debian: cleanup from Clint Byrum <clint@ubuntu.com>

With a few changes by me (changelog, copyright clarification)

15 years agofuse.cc: make license GPL
Sage Weil [Wed, 21 Jul 2010 04:25:46 +0000 (21:25 -0700)]
fuse.cc: make license GPL

Originally based on Miklos' fusexmp.c, which is GPL.  Keep that license.
LGPL makes no sense here anyway.

15 years agomon: specify default new pool size/rule/pg_num in conf
Sage Weil [Tue, 20 Jul 2010 22:02:59 +0000 (15:02 -0700)]
mon: specify default new pool size/rule/pg_num in conf

15 years agoclean up new pool crush rule values [backward compatible protocol change]
Sage Weil [Tue, 20 Jul 2010 21:59:33 +0000 (14:59 -0700)]
clean up new pool crush rule values [backward compatible protocol change]

Default to -1 (not 0!), which means 'server default'.

15 years agoosd: infer missing map is empty when last_complete==last_update
Sage Weil [Tue, 20 Jul 2010 20:24:26 +0000 (13:24 -0700)]
osd: infer missing map is empty when last_complete==last_update

15 years agomon: set/clear pause{rd,wr} osdmap flags
Sage Weil [Tue, 20 Jul 2010 20:24:08 +0000 (13:24 -0700)]
mon: set/clear pause{rd,wr} osdmap flags

15 years agoosd: clear failure_queue when marked down
Sage Weil [Tue, 20 Jul 2010 20:23:51 +0000 (13:23 -0700)]
osd: clear failure_queue when marked down

This prevents bleed through of failures (due to not getting hearbeats, due
to us being marked down) so they don't get sent after we are marked back
up again.

This fixes one possible source of up/down flapping...

15 years agorgw: fix content-length reporting
Yehuda Sadeh [Tue, 20 Jul 2010 19:12:39 +0000 (12:12 -0700)]
rgw: fix content-length reporting

and also some other various issues that came up while at it

15 years agodebug: fix log append
Sage Weil [Tue, 20 Jul 2010 17:38:26 +0000 (10:38 -0700)]
debug: fix log append

ios::ate does not work as advertised!

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agodebug: say 'append' or 'new' when opening log
Sage Weil [Tue, 20 Jul 2010 17:36:35 +0000 (10:36 -0700)]
debug: say 'append' or 'new' when opening log

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: wait for non-existant pools (w/o crashing); clean up handle_op()
Sage Weil [Tue, 20 Jul 2010 17:25:34 +0000 (10:25 -0700)]
osd: wait for non-existant pools (w/o crashing); clean up handle_op()

Make sure the pool exists before calling OSDMap::raw_pg_to_pg().  Fixes
crash on operations shortly after pg pools are created.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agorgw: fix copy
Yehuda Sadeh [Tue, 20 Jul 2010 00:20:53 +0000 (17:20 -0700)]
rgw: fix copy

15 years agorgw: get/put read and write by chunks
Yehuda Sadeh [Mon, 19 Jul 2010 23:50:43 +0000 (16:50 -0700)]
rgw: get/put read and write by chunks

15 years agorgw: lay groundwork for chunky read/write
Yehuda Sadeh [Mon, 19 Jul 2010 23:20:34 +0000 (16:20 -0700)]
rgw: lay groundwork for chunky read/write

15 years agoosd: make client op version actual at_version
Sage Weil [Mon, 19 Jul 2010 20:40:50 +0000 (13:40 -0700)]
osd: make client op version actual at_version

I.e. this should take into account any cloning.  So set it after we call
prepare_transaction, not before.

15 years agoosd: fix subop version in snaptrimmer
Sage Weil [Mon, 19 Jul 2010 20:39:18 +0000 (13:39 -0700)]
osd: fix subop version in snaptrimmer

The at_version should end equal to last_update, not +1.

15 years agomon: check command length on osdmon
Sage Weil [Mon, 19 Jul 2010 17:13:05 +0000 (10:13 -0700)]
mon: check command length on osdmon

Fixes crash on 'ceph osd tell'.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agorbd: add removal of snapshots
Yehuda Sadeh [Mon, 19 Jul 2010 17:20:01 +0000 (10:20 -0700)]
rbd: add removal of snapshots

15 years agolibrados: don't read attributes until the callback is done!
Greg Farnum [Fri, 16 Jul 2010 23:12:38 +0000 (16:12 -0700)]
librados: don't read attributes until the callback is done!

15 years agorgw: rgw_rest: Properly return the content-length.
Greg Farnum [Fri, 16 Jul 2010 19:10:16 +0000 (12:10 -0700)]
rgw: rgw_rest: Properly return the content-length.
From wido.

15 years agorgw: rgw_rados: initialize the position in list_buckets_init!
Greg Farnum [Fri, 16 Jul 2010 18:13:31 +0000 (11:13 -0700)]
rgw: rgw_rados: initialize the position in list_buckets_init!

15 years agomonc: be quiet about 'continuing hunt' too
Sage Weil [Fri, 16 Jul 2010 21:03:30 +0000 (14:03 -0700)]
monc: be quiet about 'continuing hunt' too

15 years agolibrados: librados.hpp should be standalone
Yehuda Sadeh [Fri, 16 Jul 2010 21:04:57 +0000 (14:04 -0700)]
librados: librados.hpp should be standalone

removed references to object_t, declared librados::SnapContext

15 years agomds: small EMetablob fixup
Sage Weil [Fri, 16 Jul 2010 19:03:22 +0000 (12:03 -0700)]
mds: small EMetablob fixup

Push to the back, update the back.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: always set snapset context if can_create
Sage Weil [Thu, 15 Jul 2010 21:09:13 +0000 (14:09 -0700)]
osd: always set snapset context if can_create

The most common code paths loaded it, but not the case where the obc was
already loaded.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: journal dirty items in order
Sage Weil [Fri, 16 Jul 2010 18:43:14 +0000 (11:43 -0700)]
mds: journal dirty items in order

There was some weird thing where dirty items were added to the front of
the list in the EMetaBlob, dating from 2007.  I have no idea why.  But it
was breaking rename, which dirties the same inode twice in some cases
(same src and dst dir).  Since it would replay in reverse, the inode would
end up with an older state.  If that happened to be the last time the inode
was modified in the journal prior to an mds restart+replay, we'd get bad
dirstat/rstat metadata.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: helpful journal replay debug output
Sage Weil [Fri, 16 Jul 2010 17:43:52 +0000 (10:43 -0700)]
mds: helpful journal replay debug output

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoclient: drop unnecessary #include
Sage Weil [Fri, 16 Jul 2010 17:43:37 +0000 (10:43 -0700)]
client: drop unnecessary #include

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: simplify rstat/dirstat debug printing
Sage Weil [Fri, 16 Jul 2010 17:38:57 +0000 (10:38 -0700)]
mds: simplify rstat/dirstat debug printing

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: remove bogus 'oldest snap' floor on lssnap result
Sage Weil [Thu, 15 Jul 2010 22:13:25 +0000 (15:13 -0700)]
mds: remove bogus 'oldest snap' floor on lssnap result

I suspect the intent was to exclude snaps from parents from before we
existed.  However, get_snap_info() already does that intelligently.  And
get_oldest_snap() was wrong: this dirs snaps have little to do with
the nested snaprealms, old_inodes isn't a good indicator of birth time,
etc.

This fixes a bug where we couldn't see a snap in .snaps, but after the mds
restart could (because old_inodes included more old versions after
replay).

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agolower loglevel for successfull mon hunting
Christian Brunner [Wed, 14 Jul 2010 13:43:38 +0000 (15:43 +0200)]
lower loglevel for successfull mon hunting

I'm receving log output on every rados_initialize (... found mon0).
Since this makes the qemu-iotests fail, I'd suggest to lower the loglevel.

Christian

15 years agomessages: fix MClientRequest::is_write(), rename to may_write().
Greg Farnum [Fri, 16 Jul 2010 16:56:40 +0000 (09:56 -0700)]
messages: fix MClientRequest::is_write(), rename to may_write().

Also remove unused can_forward(), auth_is_best().

15 years agomsgr: use in6addr_any extern for is_blank_addr()
Sage Weil [Thu, 15 Jul 2010 21:51:05 +0000 (14:51 -0700)]
msgr: use in6addr_any extern for is_blank_addr()

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomsgr: put brackets around ipv6 addrs
Sage Weil [Thu, 15 Jul 2010 21:49:51 +0000 (14:49 -0700)]
msgr: put brackets around ipv6 addrs

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: throw std::bad_alloc on boost:pool<> depletion
Sage Weil [Thu, 15 Jul 2010 18:38:08 +0000 (11:38 -0700)]
mds: throw std::bad_alloc on boost:pool<> depletion

A bit more informative than a segfault from this=NULL in the constructor.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agorbd: fix usage
Yehuda Sadeh [Thu, 15 Jul 2010 21:50:33 +0000 (14:50 -0700)]
rbd: fix usage

15 years agorbd: rbd tool can parse embedded snapshots name
Yehuda Sadeh [Thu, 15 Jul 2010 21:34:11 +0000 (14:34 -0700)]
rbd: rbd tool can parse embedded snapshots name

also update man page

15 years agorbd: update man page, other external tools
Yehuda Sadeh [Thu, 15 Jul 2010 21:07:52 +0000 (14:07 -0700)]
rbd: update man page, other external tools

15 years agomon: limit size of MOSDMap messages when subscribing
Sage Weil [Thu, 15 Jul 2010 18:05:16 +0000 (11:05 -0700)]
mon: limit size of MOSDMap messages when subscribing

Send multiple MOSDMap messages (max of 100 incrementals per) when
clients who are way behind subscribe.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agodebug: fix append to log file
Sage Weil [Thu, 15 Jul 2010 17:54:09 +0000 (10:54 -0700)]
debug: fix append to log file

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: oops, fix pool thing
Sage Weil [Thu, 15 Jul 2010 17:51:46 +0000 (10:51 -0700)]
osd: oops, fix pool thing

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: MOSDOp: print snapc if snap_seq != 0, not when snap list is empty
Sage Weil [Thu, 15 Jul 2010 17:43:55 +0000 (10:43 -0700)]
osd: MOSDOp: print snapc if snap_seq != 0, not when snap list is empty

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: use client or pool snapc based on _pool_ snap mode, not client op
Sage Weil [Thu, 15 Jul 2010 17:43:31 +0000 (10:43 -0700)]
osd: use client or pool snapc based on _pool_ snap mode, not client op

The pool snap mode determines where the snapc should come from.  In
particular, if the client specifies no snap seq, but the pool is in
unmanaged snap mode, then use the client (empty) snapc, and do not generate
one from the pool.

The pool might have a snap seq that is > 0, leading to and totally
invalid(!) snapc like 2=[], e.g., because the pool seq was bumped to create
an unmanaged snapc by another client.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agorgw: delete bucket objects from root pool on destruction.
Greg Farnum [Thu, 15 Jul 2010 15:23:57 +0000 (08:23 -0700)]
rgw: delete bucket objects from root pool on destruction.

This allows a bucket with that name to be created later.

15 years agorbdtool: rename rbdtool to 'rbd' tool and change usage
Yehuda Sadeh [Wed, 14 Jul 2010 23:57:30 +0000 (16:57 -0700)]
rbdtool: rename rbdtool to 'rbd' tool and change usage

15 years agocfuse: return EROFS when attempting to modify snapshots
Greg Farnum [Wed, 14 Jul 2010 20:19:29 +0000 (13:19 -0700)]
cfuse: return EROFS when attempting to modify snapshots

15 years agomds: Check all write ops to make sure they aren't in a snapshot; return EROFS if...
Greg Farnum [Tue, 13 Jul 2010 18:35:00 +0000 (11:35 -0700)]
mds: Check all write ops to make sure they aren't in a snapshot; return EROFS if they are.

15 years agorbdtool: fix add-snap
Yehuda Sadeh [Sat, 10 Jul 2010 00:07:47 +0000 (17:07 -0700)]
rbdtool: fix add-snap

15 years agorbdtool: rearrange usage
Yehuda Sadeh [Fri, 9 Jul 2010 23:23:52 +0000 (16:23 -0700)]
rbdtool: rearrange usage

15 years agorbdtool: rename should first check source
Yehuda Sadeh [Fri, 9 Jul 2010 22:50:58 +0000 (15:50 -0700)]
rbdtool: rename should first check source

that is, source image doesn't exist error should have priority
over destination already exists.

15 years agodebian: override lintian warning unstripped-binary-or-object
Sage Weil [Fri, 9 Jul 2010 22:26:03 +0000 (15:26 -0700)]
debian: override lintian warning unstripped-binary-or-object

15 years agodebug: fix debug rank symlink rotate when log_per_instance
Sage Weil [Fri, 9 Jul 2010 22:33:18 +0000 (15:33 -0700)]
debug: fix debug rank symlink rotate when log_per_instance

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agorbdtool: implement copy
Yehuda Sadeh [Fri, 9 Jul 2010 22:37:53 +0000 (15:37 -0700)]
rbdtool: implement copy

15 years agorbdtool: use canonical image name (pool/image)
Yehuda Sadeh [Fri, 9 Jul 2010 21:15:06 +0000 (14:15 -0700)]
rbdtool: use canonical image name (pool/image)

15 years agorbdtool: can read data using snapshot
Yehuda Sadeh [Fri, 9 Jul 2010 20:46:28 +0000 (13:46 -0700)]
rbdtool: can read data using snapshot

15 years agomds: fix up snap cap issue logic comments
Sage Weil [Fri, 9 Jul 2010 17:39:11 +0000 (10:39 -0700)]
mds: fix up snap cap issue logic comments

15 years agomds: fix acquire_lock() calls in handle_client_open()
Sage Weil [Fri, 9 Jul 2010 17:34:53 +0000 (10:34 -0700)]
mds: fix acquire_lock() calls in handle_client_open()

We want to acquire_locks exactly once per path through the function.
Don't look at mdr->* yet if we haven't locked yet.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: issue all snap_caps allowed given locks, not by current cap
Sage Weil [Fri, 9 Jul 2010 17:32:25 +0000 (10:32 -0700)]
mds: issue all snap_caps allowed given locks, not by current cap

For snap_caps, who cares if the live inode has a cap for this client.  We
always want to issue the maximum set of caps the inode locks allow for
a live inode or snapped inode with caps, or the max caps ever for a snapped
inode without caps.

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agofix FCGI detection and specfile
Christian Brunner [Fri, 9 Jul 2010 20:43:42 +0000 (22:43 +0200)]
fix FCGI detection and specfile

configure is building radosgw as soon as libfcgi is installed, even
if --without-radosgw is given. Here is a fix for configure.in and the
specfile.

Regards,
Christian

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomon: avoid OOB m->cmd access in auth proprocess_command
Sage Weil [Fri, 9 Jul 2010 20:14:25 +0000 (13:14 -0700)]
mon: avoid OOB m->cmd access in auth proprocess_command

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomsgr: brackets (not braces) for ipv6 addrs
Sage Weil [Fri, 9 Jul 2010 18:15:55 +0000 (11:15 -0700)]
msgr: brackets (not braces) for ipv6 addrs

Sigh

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agodebug: only truncate log file on open if log_per_instance is set
Sage Weil [Fri, 9 Jul 2010 17:45:49 +0000 (10:45 -0700)]
debug: only truncate log file on open if log_per_instance is set

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomds: termiante client sessions before flushign log during shutdown
Sage Weil [Fri, 9 Jul 2010 17:09:53 +0000 (10:09 -0700)]
mds: termiante client sessions before flushign log during shutdown

We need to close out client sessions, and caps, to empty the open_files
lists, so that we can properly trim the EOpen events from the log.

This fixes a crash like so:

mds/journal.cc: In function 'C_Gather* LogSegment::try_to_expire(MDS*)':
mds/journal.cc:160: FAILED assert(ls != this)
 1: (LogSegment::try_to_expire(MDS*)+0x11c1) [0x90a1df]
 2: (MDLog::try_expire(LogSegment*)+0x2d) [0x902c25]
 3: (MDLog::trim(int)+0x55b) [0x902b0b]
 4: (MDCache::shutdown_pass()+0xeb) [0x7d6aab]
 5: (MDS::_dispatch(Message*)+0x2dc1) [0x71597f]
 6: (MDS::ms_dispatch(Message*)+0x38) [0x712a44]
 7: (Messenger::ms_deliver_dispatch(Message*)+0x63) [0x6fe07b]
 8: (SimpleMessenger::dispatch_entry()+0x5fd) [0x6ef7e7]
 9: (SimpleMessenger::DispatchThread::entry()+0x2c) [0x6e44b8]
 10: (Thread::_entry_func(void*)+0x23) [0x6fd075]
 11: /lib64/libpthread.so.0() [0x3552e06a3a]
 12: (clone()+0x6d) [0x35526ddf3d]
 NOTE: a copy of the executable, or `objdump -rdS <executable>` is
needed to interpret this.

Reported-by: Henry C Chang <henry_c_chang@tcloudcomputing.com>
Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoceph.spec.in: create /var/log/ceph directory
Christian Brunner [Fri, 9 Jul 2010 14:01:47 +0000 (16:01 +0200)]
ceph.spec.in: create /var/log/ceph directory

I've missed this one in my previous patch series.

Regards,
Christian

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoceph.spec.in: add a build option for radosgw (--with radosgw)
Christian Brunner [Fri, 9 Jul 2010 09:13:04 +0000 (11:13 +0200)]
ceph.spec.in: add a build option for radosgw (--with radosgw)

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoceph.spec.in: add /var/lib/ceph/
Christian Brunner [Fri, 9 Jul 2010 09:12:39 +0000 (11:12 +0200)]
ceph.spec.in: add /var/lib/ceph/

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoceph.spec.in: add logrotate configuration
Christian Brunner [Fri, 9 Jul 2010 09:12:16 +0000 (11:12 +0200)]
ceph.spec.in: add logrotate configuration

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoceph.spec.in: set correct path for rbd classes
Christian Brunner [Fri, 9 Jul 2010 09:11:49 +0000 (11:11 +0200)]
ceph.spec.in: set correct path for rbd classes

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agoosd: print debug write before adding to transaction
Sage Weil [Thu, 8 Jul 2010 22:41:32 +0000 (15:41 -0700)]
osd: print debug write before adding to transaction

Signed-off-by: Sage Weil <sage@newdream.net>
15 years agomsgr: return error and dethrottle if message decode fails
Sage Weil [Thu, 8 Jul 2010 21:55:53 +0000 (14:55 -0700)]
msgr: return error and dethrottle if message decode fails

This was broken by the addition of the throttling stuff.

Signed-off-by: Sage Weil <sage@newdream.net>