]>
git.apps.os.sepia.ceph.com Git - ceph.git/log
Samuel Just [Thu, 15 Nov 2012 00:30:51 +0000 (16:30 -0800)]
msg/DispatchQueue: release throttle on messages when dropping an id
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Thu, 15 Nov 2012 00:30:05 +0000 (16:30 -0800)]
PrioritizedQueue: allow remove_by_class to return removed items
Signed-off-by: Samuel Just <sam.just@inktank.com>
Sage Weil [Thu, 15 Nov 2012 00:59:06 +0000 (16:59 -0800)]
librbd: use delete[] properly
==4986== Mismatched free() / delete / delete []
==4986== at 0x4C2658C: operator delete(void*) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==4986== by 0x4ED8EA9: librbd::ImageCtx::~ImageCtx() (ImageCtx.cc:100)
==4986== by 0x4EF3827: librbd::close_image(librbd::ImageCtx*) (internal.cc:1869)
==4986== by 0x4EE8FB8: librbd::clone(librados::IoCtx&, char const*, char const*, librados::IoCtx&, char const*, unsigned long, int*, unsigned long, int) (internal.cc:900)
==4986== by 0x4EC363C: rbd_clone2 (librbd.cc:553)
==4986== by 0x404C85: do_clone (fsx.c:836)
==4986== by 0x405639: test (fsx.c:1048)
==4986== by 0x406369: main (fsx.c:1523)
==4986== Address 0xd498b30 is 0 bytes inside a block of size 37 alloc'd
==4986== at 0x4C26CF7: operator new[](unsigned long) (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==4986== by 0x4ED9B4D: librbd::ImageCtx::init_layout() (ImageCtx.cc:164)
==4986== by 0x4ED9845: librbd::ImageCtx::init() (ImageCtx.cc:142)
==4986== by 0x4EF3449: librbd::open_image(librbd::ImageCtx*, bool) (internal.cc:1828)
==4986== by 0x4EE89E0: librbd::clone(librados::IoCtx&, char const*, char const*, librados::IoCtx&, char const*, unsigned long, int*, unsigned long, int) (internal.cc:871)
==4986== by 0x4EC363C: rbd_clone2 (librbd.cc:553)
==4986== by 0x404C85: do_clone (fsx.c:836)
==4986== by 0x405639: test (fsx.c:1048)
==4986== by 0x406369: main (fsx.c:1523)
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 15 Nov 2012 00:54:17 +0000 (16:54 -0800)]
objecter: fix leak of out_handlers
The error paths don't use the handlers. Make sure they get cleaned up.
Fixes: #3446
Signed-off-by: Sage Weil <sage@inktank.com>
Yehuda Sadeh [Thu, 15 Nov 2012 00:42:11 +0000 (16:42 -0800)]
rgw: update post policy parser
json parser semantics changed a little bit, so
needed to update the policy parser.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sage Weil [Thu, 15 Nov 2012 00:26:58 +0000 (16:26 -0800)]
mon: set default port when binding to random local ip
Fixes #3135
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Thu, 15 Nov 2012 00:22:27 +0000 (16:22 -0800)]
Merge remote-tracking branch 'gh/wip-asok' into next
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
Yehuda Sadeh [Wed, 14 Nov 2012 19:30:34 +0000 (11:30 -0800)]
rgw: relax date format check
Don't try to parse beyond the GMT or UTC. Some clients use
special date formatting. If we end up misparsing the date
it'll fail in the authorization, so don't need to be too
restrictive.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sage Weil [Wed, 14 Nov 2012 02:18:24 +0000 (18:18 -0800)]
client: register admin socket commands without lock held
Avoid a lock cycle.
existing dependency Client::client_lock (11) -> AdminSocket::m_lock (16) at:
ceph version
0.54-578-g7926ef5 (
7926ef53935313501d4a7fe0e587f3e3b00b313c )
1: (Mutex::Lock(bool)+0x41) [0x831337]
2: (AdminSocket::register_command(std::string, AdminSocketHook*, std::string)+0x40) [0x873a32]
3: (Client::init()+0x454) [0x6f4c24]
4: (main()+0x637) [0x6ea399]
5: (__libc_start_main()+0xed) [0x7fd97bbca76d]
6: ./ceph-fuse() [0x6e9c59]
-4> 2012-11-13 18:14:48.619714
7fd97b1a3700 0 new dependency AdminSocket::m_lock (16) -> Client::client_lock (11) creates a cycle at
ceph version
0.54-578-g7926ef5 (
7926ef53935313501d4a7fe0e587f3e3b00b313c )
1: (Mutex::Lock(bool)+0x41) [0x831337]
2: (Objecter::RequestStateHook::call(std::string, std::string, ceph::buffer::list&)+0x7a) [0x90627e]
3: (AdminSocket::do_accept()+0xb1b) [0x87318f]
4: (AdminSocket::entry()+0x2fa) [0x8725fe]
5: (Thread::_entry_func(void*)+0x23) [0x86b335]
6: (()+0x7e9a) [0x7fd97d279e9a]
7: (clone()+0x6d) [0x7fd97bc9ccbd]
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Wed, 14 Nov 2012 02:17:55 +0000 (18:17 -0800)]
objecter: separate locked and unlocked init/shutdown
We don't want to hold the lock while we register the admin socket commands
or else we create a lock cycle when we try to process them later.
Signed-off-by: Sage Weil <sage@inktank.com>
Gary Lowell [Wed, 14 Nov 2012 01:29:47 +0000 (17:29 -0800)]
Merge branch 'next'
Conflicts:
configure.ac
src/rgw/rgw_common.cc
Samuel Just [Wed, 14 Nov 2012 00:45:49 +0000 (16:45 -0800)]
osd/: add config helper for min_size and update build_simple*
min_size should never be set to 0 on a pool. config.h
now has a helper to determine the correct default value.
Signed-off-by: Samuel Just <sam.just@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>
Sage Weil [Wed, 14 Nov 2012 01:11:34 +0000 (17:11 -0800)]
doc/release-notes: fix heading
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Wed, 14 Nov 2012 00:24:23 +0000 (16:24 -0800)]
doc: release-notes for v0.54
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 13 Nov 2012 22:34:53 +0000 (14:34 -0800)]
doc: update crush weight ramping process
Signed-off-by: Sage Weil <sage@inktank.com>
Yehuda Sadeh [Tue, 13 Nov 2012 23:42:52 +0000 (15:42 -0800)]
rgw: fix warning
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sage Weil [Tue, 13 Nov 2012 23:39:42 +0000 (15:39 -0800)]
Merge branch 'wip-min-size'
Reviewed-by: Sam Just <sam.just@inktank.com>
Sage Weil [Tue, 13 Nov 2012 23:16:56 +0000 (15:16 -0800)]
osd: default pool min_size to 0 (which gives us size-size/2)
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 13 Nov 2012 23:11:42 +0000 (15:11 -0800)]
mon: default min_size to size-size/2 if min_size default is 0
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 13 Nov 2012 21:27:45 +0000 (13:27 -0800)]
osd: default min_size to size - size/2
size -> min_size:
5 -> 3
4 -> 2
3 -> 2
2 -> 1
Basically, default to tolerating minority down.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 13 Nov 2012 21:25:50 +0000 (13:25 -0800)]
mon: helpful warning in 'health detail' output about incomplete pgs
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Tue, 13 Nov 2012 21:25:31 +0000 (13:25 -0800)]
osd: start_boot() after init()
The previous trigger for start_boot() was racy, depending on whether we
got our rotating keys quickly.
Signed-off-by: Sage Weil <sage@inktank.com>
Dan Mick [Tue, 13 Nov 2012 21:24:15 +0000 (13:24 -0800)]
vstart.sh: support -X by adding 'auth required = none' entries
Signed-off-by: Dan Mick <dan.mick@inktank.com>
Sage Weil [Tue, 13 Nov 2012 22:50:42 +0000 (14:50 -0800)]
Merge remote-tracking branch 'gh/wip-rgw-integration'
Conflicts:
src/common/config_opts.h
Gary Lowell [Tue, 13 Nov 2012 21:18:07 +0000 (13:18 -0800)]
v0.54
Yehuda Sadeh [Tue, 13 Nov 2012 21:06:22 +0000 (13:06 -0800)]
rgw: compile with -Woverloaded-virtual
This will trigger a warning if RGWRados api changes while
RGWCache doesn't.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Tue, 13 Nov 2012 20:09:05 +0000 (12:09 -0800)]
rgw: fix RGWCache api
RGWCache api diverted form RGWRados, crippling the cache.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Yehuda Sadeh [Tue, 13 Nov 2012 20:09:05 +0000 (12:09 -0800)]
rgw: fix RGWCache api
RGWCache api diverted form RGWRados, crippling the cache.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sage Weil [Mon, 12 Nov 2012 15:06:42 +0000 (07:06 -0800)]
osd: remove dead rotating key code from init
Ancient, dead.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 15:06:25 +0000 (07:06 -0800)]
osd: defer boot until we have rotating keys
Make sure we have our rotating keys before we start booting. This
ensures we can open connections with peers *before* we add ourselves to
the osdmap. This behaviors marks instances of #3292, although it is
not clear whether it is responsible for the actual crash.
Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Sam Just <sam.just@inktank.com>
Samuel Just [Tue, 13 Nov 2012 18:56:22 +0000 (10:56 -0800)]
Merge branches 'wip_persist_missing' and 'wip_recovery_qos'
Reviewed-by: Sage Weil <sage@inktank.com>
Samuel Just [Mon, 5 Nov 2012 23:40:43 +0000 (15:40 -0800)]
PG: persist divergent_priors in ondisklog
Consider the following logs:
a) 10'10(5'7) foo
12'11(4'3) bar
b) 10'10(5'7) foo
13'11(4'4) baz
When the osd with a merges primary log b, bar is deleted and
added to the missing set with need=4'3 and have=0'0. If
the osd then dies after deleting bar, but before recovering
bar, PG::read_state() on start up will fail to re-add bar
to the missing set, and bar will be incorrect on that osd.
Now, (4'3, bar) will be added to the divergent_priors mapping
to be scanned during read_state along with the log.
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Mon, 5 Nov 2012 19:33:13 +0000 (11:33 -0800)]
PG::merge_old_entry: fix case for divergent prior_version
Previously, we asserted that a log entry with a divergent
prior_version must be a clone. Consider the following
case:
6'11(6'2) m foo
7'12(6'3) m bar
7'13(7'12) m bar
If this is merged with:
6'11(6'2) m foo
8'12(6'4) m baz
we will hit the assert.
Merging a divergent entry with prior_version after current
tail, but not in the log implies that prior_version was a
divergent entry which we have already merged. The missing
set and filestore collection must therefore have already
been adjusted.
Signed-off-by: Samuel Just <sam.just@inktank.com>
Sage Weil [Sat, 10 Nov 2012 11:57:23 +0000 (03:57 -0800)]
PrioritizedQueue: use iterator to streamlink SubQueue::remove_by_class()
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 10 Nov 2012 11:51:25 +0000 (03:51 -0800)]
PrioritizedQueue: avoid double-lookup on create_queue()
Signed-off-by: Sage Weil <sage@inktank.com>
Samuel Just [Sat, 29 Sep 2012 00:27:39 +0000 (17:27 -0700)]
osd/: de-prioritize recovery ops relative to client ops
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Tue, 2 Oct 2012 23:20:24 +0000 (16:20 -0700)]
msg/: use PrioritizedQueue to handle DispatchQueue queueing
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Sat, 29 Sep 2012 00:24:40 +0000 (17:24 -0700)]
OSD: queue ops based on message priority
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Mon, 1 Oct 2012 23:11:40 +0000 (16:11 -0700)]
OSD: use PrioritizedQueue for OpWQ
The OpWQ PriorityQueue replaces OSD::op_queue, PG::op_queue,
and PG::qlock. The syncronization is now done as part of the
usual WorkQueue syncronization pattern.
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Fri, 28 Sep 2012 18:50:59 +0000 (11:50 -0700)]
WorkQueue: Add WorkQueueVal for passing params by val
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Fri, 28 Sep 2012 18:50:30 +0000 (11:50 -0700)]
common: Add PriorititizedQueue
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Tue, 2 Oct 2012 23:34:50 +0000 (16:34 -0700)]
Pipe: we don't queue DispatchQueue codes on an actual pipe
The dispatch queue event codes are now queued directly and
won't end up in a real Pipe.
Signed-off-by: Samuel Just <sam.just@inktank.com>
Samuel Just [Tue, 2 Oct 2012 20:33:11 +0000 (13:33 -0700)]
common: add RefCountedObj.cc with intrusive_ptr hooks
Signed-off-by: Samuel Just <sam.just@inktank.com>
caleb miles [Fri, 26 Oct 2012 19:17:05 +0000 (15:17 -0400)]
rgw_cache: change call signature to overwrite rgw_rados put_obj_meta()
Signed-off-by: caleb miles <caleb.miles@inktank.com>
Yehuda Sadeh [Wed, 7 Nov 2012 23:39:56 +0000 (15:39 -0800)]
rgw: unquote etag for if-match
Fixes: #3453
Passed ETag for copy/get request may be quoted.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Sylvain Munaut [Thu, 8 Nov 2012 20:11:59 +0000 (12:11 -0800)]
rgw: use thread safe version of gmtime
Use gmtime_r instead of gmtime.
Signed-off-by: Sylvain Munaut <s.munaut@whatever-company.com>
Yehuda Sadeh [Wed, 7 Nov 2012 21:21:15 +0000 (13:21 -0800)]
rgw: don't convert object mtime to UTC
Fixes: #3452
When we read object info, don't try to convert mtime to
UTC, it's already in UTC.
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
Noah Watkins [Tue, 13 Nov 2012 17:39:20 +0000 (09:39 -0800)]
java: remove redundant src.dir
Previous wip-fix-javadocs had an oversight, redefining src.dir.
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
Josh Durgin [Tue, 13 Nov 2012 16:14:09 +0000 (08:14 -0800)]
doc: reorder some openstack/rbd instructions
Move client creation to the section on setting up client auth, so you
don't skip it if you already have pools created.
Move CEPH_ARGS setting to the section on configuring OpenStack, since
it's a change for the OpenStack services, not purely ceph client
configuration.
A couple people were confused by the placement of these parts, and
they make more sense in these sections.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Tue, 13 Nov 2012 01:01:30 +0000 (17:01 -0800)]
Merge branch 'wip-client-asok'
Noah Watkins [Mon, 12 Nov 2012 23:49:38 +0000 (15:49 -0800)]
java: fix build.xml formatting
set noet ts=2 sw=2 sws=2
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
Noah Watkins [Mon, 12 Nov 2012 23:44:06 +0000 (15:44 -0800)]
java: fix javadoc builds
Don't build JavaDoc for tests, and fix the missing src.dir variable.
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
Sage Weil [Mon, 12 Nov 2012 23:40:08 +0000 (15:40 -0800)]
osdc/ObjectCacher: only return ENOENT if ObjectSet is flagged
The fs client can't handle ENOENT from the cache, but librbd wants it.
Also, the fs client will send down multiple ObjectExtents per io, but that
is incompatible with the ENOENT behavior.
Indicate which behavior we want via the ObjectSet, and update librbd to
explicitly ask for it. This fixes the fs client, which is currently
broken (it returns ENOENT on read).
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 23:00:11 +0000 (15:00 -0800)]
client: unregister commands on shutdown
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 20:02:17 +0000 (12:02 -0800)]
client: fix null dentry crash on dump_cache
Dentries can be NULL!
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 20:01:50 +0000 (12:01 -0800)]
client: dump mds session info
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 19:06:42 +0000 (11:06 -0800)]
client: add dump_cache asok command
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 19:03:33 +0000 (11:03 -0800)]
common: dumpers for ceph_{file,dir}_layout
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 04:37:56 +0000 (20:37 -0800)]
client: add mds_requests asok command
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 10 Nov 2012 09:46:33 +0000 (01:46 -0800)]
Makefile: fix hadoop lib build
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 05:12:39 +0000 (21:12 -0800)]
Makefile: use libclient.la for libcephfs
Avoid building these files twice!
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 23:03:12 +0000 (15:03 -0800)]
Merge branch 'next'
Sage Weil [Mon, 12 Nov 2012 23:02:51 +0000 (15:02 -0800)]
Merge branch 'wip-client-leaks' into next
Sage Weil [Mon, 12 Nov 2012 23:02:41 +0000 (15:02 -0800)]
client: fix null put in ~MetaSession
Signed-off-by: Sage Weil <sage@inktank.com>
Josh Durgin [Mon, 12 Nov 2012 21:59:36 +0000 (13:59 -0800)]
librbd: fix create existence checking
cda9e516b8bb09b8846814cc8d4ee2879a53b2d5 made us return 0 when the
image already existed, causing copy to erroneosly ignore an existing
image. Separate the case where we know the image exists from being
unable to tell whether it exists because of e.g. an authentication
problem.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Mon, 12 Nov 2012 21:57:18 +0000 (13:57 -0800)]
librbd: debug when copy occurs
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Mon, 12 Nov 2012 20:36:19 +0000 (12:36 -0800)]
Merge remote-tracking branch 'gh/wip-javadoc'
Sage Weil [Mon, 12 Nov 2012 20:30:14 +0000 (12:30 -0800)]
Merge remote-tracking branch 'gh/wip-librbd-remove-cleanup'
Sage Weil [Mon, 22 Oct 2012 18:29:38 +0000 (11:29 -0700)]
osd: add 'osd debug drop op probability'
This is meant to exercise the kclient's timeout and osd reconnect logic
by dropping some client requests on the floor.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 20:01:07 +0000 (12:01 -0800)]
mon: require pgnum in 'ceph osd pool create <poolname> <pgnum> [<pgp_num>]' command
The default of 8 is virtually never the right answer. Require the initial
pg count to be explicitly provided.
Signed-off-by: Sage Weil <sage@inktank.com>
Josh Durgin [Thu, 8 Nov 2012 00:48:32 +0000 (16:48 -0800)]
librbd: return actual error when detecting format fails during creation
This bit a couple users today, when bad osd caps resulted in a very
confusing error message.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Mon, 12 Nov 2012 18:41:30 +0000 (10:41 -0800)]
Merge remote branch 'origin/wip-rbd-read'
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Mon, 12 Nov 2012 19:08:46 +0000 (11:08 -0800)]
Merge remote-tracking branch 'gh/wip-client-symlinks'
Reviewed-by: Sam Lang <sam.lang@inktank.com>
Noah Watkins [Wed, 7 Nov 2012 04:17:25 +0000 (20:17 -0800)]
java: add symlink/readlink tests
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
Sage Weil [Mon, 12 Nov 2012 18:07:49 +0000 (10:07 -0800)]
test_libcephfs: fix, add symlink loop tests
The first test did
/a/b/file
/a/b/sym -> /a/b
and opened /a/b/sym/file, which is valid. Change it to
/a/b/file
/a/b/sym -> /a/b/sym
which is not.
Add another test that does
/a -> /b
/b -> /c
/c -> /a
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Mon, 12 Nov 2012 16:56:45 +0000 (08:56 -0800)]
debug: adjust default debug levels
Trim out most noise, keep things that are interesting.
Notably, we are logging each message sent and received, and we are logging
the filestore operations when they get queued. Those may still benefit
from being turned off in high IOPS environments.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 10 Nov 2012 10:40:33 +0000 (02:40 -0800)]
Merge branch 'next'
Sage Weil [Sat, 10 Nov 2012 10:38:26 +0000 (02:38 -0800)]
Merge branch 'wip-client-leaks' into next
Sage Weil [Sat, 10 Nov 2012 10:35:04 +0000 (02:35 -0800)]
client: simplify/fix symlink loop check
Checking that we visit a symlink isn't correct; for example, the below is
valid, and we visit /b twice.
/a/b -> c
/a/c/d -> /a/b
In order to do this "correctly", I think we would need to track the pairs
of paths and symlinks we are resolving. But, reading the man pages,
ELOOP is actually just defined as traversing more than MAXSYMLINKS syms.
(It appears to be 20 on my machine.)
Just do that instead.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Sat, 10 Nov 2012 10:14:02 +0000 (02:14 -0800)]
client: fix path_walk for directory symlinks
Signed-off-by: Sage Weil <sage@inktank.com>
Josh Durgin [Thu, 8 Nov 2012 02:19:07 +0000 (18:19 -0800)]
OSDMonitor: remove max_devices and max_osd interdependency
Higher max_osd than max_devices doesn't hurt anything (and is the
normal way to add more osds). Higher max_devices than max_osds are
filtered out of crush results since
e541c0f8d871172ec61962372efca943308e5fe ,
so they don't matter either.
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Fri, 9 Nov 2012 13:28:12 +0000 (05:28 -0800)]
mds: re-try_set_loner() after doing evals in eval(CInode*, int mask)
Consider a case where current loner is A and wanted loner is B.
At the top of the function we try to set the loner, but that may fail
because we haven't processed the gathered caps yet for the previous
loner. In the body we do that and potentially drop the old loner, but we
do not try_set_loner() again on the desired loner.
Try after our drop. If it succeeds, loop through the eval's one more time
so that we can issue caps approriately.
This fixes a hang induced by a simple loop like:
while true ; do echo asdf >> mnt.a/foo ; tail mnt.b/foo ; done &
while true ; do ls mnt.a mnt.b ; done
(The second loop may not be necessary.)
Signed-off-by: Sage Weil <sage@inktank.com>
Dan Mick [Sat, 10 Nov 2012 03:35:12 +0000 (19:35 -0800)]
Merge branch 'next'
Pull in types.h fix
Dan Mick [Sat, 10 Nov 2012 03:17:56 +0000 (19:17 -0800)]
si_t was not properly converting values < 100KB
Signed-off-by: Dan Mick <dan.mick@inktank.com>
Sage Weil [Sat, 10 Nov 2012 02:34:22 +0000 (18:34 -0800)]
osdc/ObjectCacher: only call flush callback if we transitions to clean
If we race with e.g. truncate and are in bh_write_commit but the oset
is already clean, we should not call the flush callback (again).
This is reproduced by:
- kludging slow osd replies into the code (e.g., 2 second delay)
- mount ceph-fuse with --client-oc-max-dirty-age 1
- dd if=/dev/zero of=mnt/foo count=1
sleep 1
truncate --size 0 mnt/foo
-> crash
Signed-off-by: Sage Weil <sage@inktank.com>
tamil [Sat, 10 Nov 2012 01:55:20 +0000 (17:55 -0800)]
cleaned up scripts
Signed-off-by: tamil <tamil.muthamizhan@inktank.com>
Sage Weil [Fri, 9 Nov 2012 18:25:48 +0000 (10:25 -0800)]
client: ensure we don't leak MClientReply
We are careful to clear this reference when processing it.
Add an assert here. There's no way we can get 2 quick replies because
of the kick-back below.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 19:32:44 +0000 (11:32 -0800)]
ceph-fuse: fix leak of args
Also fix up the helper we use to have fewer sharp edges.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 19:02:51 +0000 (11:02 -0800)]
osdc/ObjectCacher: fix leak on readahead
If we initiate io (success == false) but have no waiter, we need to
delete the OSDRead.
This affects libcephfs/ceph-fuse, but not librbd, which does no readahead.
Signed-off-by: Sage Weil <sage@inktank.com>
Gary Lowell [Fri, 9 Nov 2012 21:28:13 +0000 (13:28 -0800)]
ceph.spec.in: Build debuginfo subpackage.
This is a partial fix for bug 3471. Enable building of debuginfo package.
Some distributions enable this automatically by installing additional rpm
macros, on others it needs to be explicity added to the spec file.
Josh Durgin [Fri, 9 Nov 2012 20:17:24 +0000 (12:17 -0800)]
test: add cli test for missing args to rbd
This includes 'rbd mv foo', which used to crash
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 9 Nov 2012 20:11:18 +0000 (12:11 -0800)]
rbd: check for second argument to mv/rename
Without this check, 'rbd mv foo' crashed trying to use a NULL char* as
a string.
Reported-by: Andrey Korolyov <andrey@xdel.ru>
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Josh Durgin [Fri, 9 Nov 2012 20:11:18 +0000 (12:11 -0800)]
rbd: check for second argument to mv/rename
Without this check, 'rbd mv foo' crashed trying to use a NULL char* as
a string.
Reported-by: Andrey Korolyov <andrey@xdel.ru>
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Sam Lang [Fri, 9 Nov 2012 19:52:07 +0000 (13:52 -0600)]
Merge branch 'wip-java-test'
Joe Buck [Thu, 8 Nov 2012 18:09:24 +0000 (10:09 -0800)]
test.sh: Adding LD_LIBRARY_PATH so JUnit can find .so files.
Removing useless java.library.path references.
Signed-off-by: Joe Buck <jbbuck@gmail.com>
Reviewed-by: Noah Watkins <noahwatkins@gmail.com>
Samuel Just [Fri, 9 Nov 2012 19:24:42 +0000 (11:24 -0800)]
doc/cluster-ops/pools.rst: Added documentation for min_size
Signed-off-by: Samuel Just <sam.just@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
Sage Weil [Fri, 9 Nov 2012 18:12:21 +0000 (10:12 -0800)]
client: fix SnapRealm leak
Start ref count at 0; get_snap_realm() will increment it after alloc.
Fix the ref drop order so that the xlist is empty.
Signed-off-by: Sage Weil <sage@inktank.com>
Sage Weil [Fri, 9 Nov 2012 18:09:53 +0000 (10:09 -0800)]
client: debug SnapRealm reference counting
Signed-off-by: Sage Weil <sage@inktank.com>
Noah Watkins [Fri, 9 Nov 2012 04:16:53 +0000 (20:16 -0800)]
vstart: allow minimum pool size of one
A minimum pool size of 2 was introduced by
13486857cf . This sets the
minimum to one so that basic vstart environments work.
Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
Reviewed-by: Samuel Just <sam.just@inktank.com>
Sage Weil [Fri, 9 Nov 2012 18:02:59 +0000 (10:02 -0800)]
client: fix leak of Cap
Signed-off-by: Sage Weil <sage@inktank.com>