]> git.apps.os.sepia.ceph.com Git - ceph.git/log
ceph.git
13 years agoceph: add a new "run_uml.sh" script to manage running a UML client
Alex Elder [Tue, 10 Jan 2012 02:13:41 +0000 (18:13 -0800)]
ceph: add a new "run_uml.sh" script to manage running a UML client

This script is used to automate most of what's required to run a
User-Mode Linux (UML) instance.  This is mainly of interest for
ceph client developers who might benefit from the debugger access
that UML affords.  It was written for ceph development but isn't
really dependent on ceph.  It basically makes a few assumptions and
follows some conventions, and in doing so is able to encapsulate
most of the "tricky parts" of setting up to run a UML instance.

Signed-off-by: Alex Elder <elder@dreamhost.com>
13 years agorgw: adjust log level
Yehuda Sadeh [Mon, 9 Jan 2012 19:40:44 +0000 (11:40 -0800)]
rgw: adjust log level

13 years agorgw: some cleanup
Yehuda Sadeh [Mon, 9 Jan 2012 18:31:15 +0000 (10:31 -0800)]
rgw: some cleanup

13 years agorgw: only use plain PUT processor when !chunked_upload
Yehuda Sadeh [Mon, 9 Jan 2012 18:15:00 +0000 (10:15 -0800)]
rgw: only use plain PUT processor when !chunked_upload

13 years agoMerge remote branch 'gh/wip-osd-retry-attempt'
Sage Weil [Sun, 8 Jan 2012 18:15:41 +0000 (10:15 -0800)]
Merge remote branch 'gh/wip-osd-retry-attempt'

13 years agoMerge remote branch 'gh/wip-admin-socket'
Sage Weil [Sun, 8 Jan 2012 16:16:56 +0000 (08:16 -0800)]
Merge remote branch 'gh/wip-admin-socket'

13 years agoperfcounters: fix unittest for new admin_socket interface
Sage Weil [Sat, 7 Jan 2012 03:09:10 +0000 (19:09 -0800)]
perfcounters: fix unittest for new admin_socket interface

Broken by b389685afa1be00b5147855bf71c50042bfbfa6c.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
13 years agoMakefile: disable untitest_interval_tree
Sage Weil [Sat, 7 Jan 2012 04:39:05 +0000 (20:39 -0800)]
Makefile: disable untitest_interval_tree

Segfaults. Valgrind errors. Accessing uninitialized memory.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
13 years agounittest_interval_tree: make it compile
Sage Weil [Sat, 7 Jan 2012 04:38:33 +0000 (20:38 -0800)]
unittest_interval_tree: make it compile

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
13 years agoosd: read op should claim_append data instead of claim
Yehuda Sadeh [Sat, 7 Jan 2012 00:55:52 +0000 (16:55 -0800)]
osd: read op should claim_append data instead of claim

13 years agorgw: remove object before writing both xattrs and data
Yehuda Sadeh [Sat, 7 Jan 2012 00:51:23 +0000 (16:51 -0800)]
rgw: remove object before writing both xattrs and data

otherwise we'll leak xattrs from previous incarnation

13 years agorgw: create plain processor for small objects
Yehuda Sadeh [Sat, 7 Jan 2012 00:16:06 +0000 (16:16 -0800)]
rgw: create plain processor for small objects

13 years agorgw: fix multipart PUT
Yehuda Sadeh [Fri, 6 Jan 2012 23:07:08 +0000 (15:07 -0800)]
rgw: fix multipart PUT

latest revamp broke it, missed calling RGWPutObjProcessor::prepare(s)
where needed.

13 years agorgw: rearrange PutObj::execute()
Yehuda Sadeh [Fri, 6 Jan 2012 20:41:33 +0000 (12:41 -0800)]
rgw: rearrange PutObj::execute()

groundwork for different handling of small object PUTs

13 years agorgw: different atomic handling for small objects
Yehuda Sadeh [Thu, 5 Jan 2012 20:51:27 +0000 (12:51 -0800)]
rgw: different atomic handling for small objects

13 years agomon: fix uninitialized cluster_logger_registered
Sage Weil [Fri, 6 Jan 2012 22:32:16 +0000 (14:32 -0800)]
mon: fix uninitialized cluster_logger_registered

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoobjecter: ignore replies from old request attempts
Sage Weil [Fri, 6 Jan 2012 19:38:15 +0000 (11:38 -0800)]
objecter: ignore replies from old request attempts

If we know the request attempt, ignore old attempts.

If we do not know the attempt (because the server is old), accept the
reply.  This could lead to doing some ACK callbacks we shouldn't in
extreme failure/recovery scenarios, but that is better than doing
the callbacks out of order.

Partially fixes: #1490
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosd: encode retry attempt in MOSDOp[Reply]
Sage Weil [Fri, 6 Jan 2012 20:49:13 +0000 (12:49 -0800)]
osd: encode retry attempt in MOSDOp[Reply]

In addition to the boolean flag, also encode the exact retry attempt.

Return -1 if we don't know.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: document quorum_status, mon_status
Sage Weil [Fri, 6 Jan 2012 20:20:18 +0000 (12:20 -0800)]
mon: document quorum_status, mon_status

Fixes: #1824
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: fix misplaced else
Sage Weil [Fri, 6 Jan 2012 20:19:59 +0000 (12:19 -0800)]
mon: fix misplaced else

Broken by 435c29448a10ec343f5a2b7195d94c72de5b1a25.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoMerge remote branch 'gh/wip-mon-timeouts'
Sage Weil [Fri, 6 Jan 2012 18:20:55 +0000 (10:20 -0800)]
Merge remote branch 'gh/wip-mon-timeouts'

13 years agoceph: speak new admin socket protocol
Sage Weil [Thu, 5 Jan 2012 21:58:05 +0000 (13:58 -0800)]
ceph: speak new admin socket protocol

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoadmin_socket: fix, extend admin_socket unit tests
Sage Weil [Fri, 6 Jan 2012 17:30:54 +0000 (09:30 -0800)]
admin_socket: fix, extend admin_socket unit tests

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoadmin_socket: string commands
Sage Weil [Thu, 5 Jan 2012 21:57:58 +0000 (13:57 -0800)]
admin_socket: string commands

Commands are strings.  Old __be32 works too.  'help' to list available
commands.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: elector needs to reset leader_acked on every election start
Greg Farnum [Thu, 5 Jan 2012 23:29:32 +0000 (15:29 -0800)]
mon: elector needs to reset leader_acked on every election start

Otherwise you never reset the leader_acked after a failed
election attempt, so if mon 0 is available on the first round
but then fails, you never make progress!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agomon: instrument elector so you can stop participating in the quorum
Greg Farnum [Thu, 5 Jan 2012 23:36:37 +0000 (15:36 -0800)]
mon: instrument elector so you can stop participating in the quorum

Add new monitor commands "quorum exit" and "quorum enter" to use it.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agomon: kill client sessions when we're not in quorum
Greg Farnum [Thu, 5 Jan 2012 22:03:43 +0000 (14:03 -0800)]
mon: kill client sessions when we're not in quorum

After a timeout of 2*mon_lease length (ie, two election rounds),
kill existing client sessions so they can reconnect to a
monitor that's (hopefully) remained in the quorum. Let any
new client sessions stick around for a mon_lease interval, then
do the same to them.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agoOCF RA: fix variable name
Florian Haas [Thu, 5 Jan 2012 21:33:32 +0000 (22:33 +0100)]
OCF RA: fix variable name

13 years agodebian: build ceph-resource-agents
Florian Haas [Thu, 5 Jan 2012 21:33:31 +0000 (22:33 +0100)]
debian: build ceph-resource-agents

13 years agoadmin_socket: refactor
Sage Weil [Thu, 5 Jan 2012 18:46:31 +0000 (10:46 -0800)]
admin_socket: refactor

Combine AdminSocketConfigObs with AdminSocket so that we can interact
with it via the cct.  Simpler class structure.  Less pointer indirection.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agorbd: add a command to delete all snapshots of an image
Josh Durgin [Thu, 5 Jan 2012 01:07:07 +0000 (17:07 -0800)]
rbd: add a command to delete all snapshots of an image

This makes deleting images with many snapshots easier.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoadmin_socket: whitespace
Sage Weil [Thu, 5 Jan 2012 17:32:49 +0000 (09:32 -0800)]
admin_socket: whitespace

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agolibrbd: don't remove an image that still has snapshots
Josh Durgin [Thu, 5 Jan 2012 01:05:49 +0000 (17:05 -0800)]
librbd: don't remove an image that still has snapshots

Return -EBUSY instead. After the header is removed, the snapshots
can't be removed or read, so make sure they're gone before proceeding.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoSimpleMessenger: clarify when ms_bind_ipv6 is used
Josh Durgin [Thu, 5 Jan 2012 01:34:17 +0000 (17:34 -0800)]
SimpleMessenger: clarify when ms_bind_ipv6 is used

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoqa: add a slightly more stressful anchortable test
Greg Farnum [Thu, 5 Jan 2012 01:08:45 +0000 (17:08 -0800)]
qa: add a slightly more stressful anchortable test

This creates more than 8 links.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agoqa: fix mdstable script for proper injectargs use.
Greg Farnum [Wed, 4 Jan 2012 23:36:08 +0000 (15:36 -0800)]
qa: fix mdstable script for proper injectargs use.

This script is fairly primitive, but somebody might find it useful...

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agorados: fix run-length option parsing for rados load-gen
Sage Weil [Thu, 5 Jan 2012 00:25:29 +0000 (16:25 -0800)]
rados: fix run-length option parsing for rados load-gen

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>
13 years agoMerge pull request #8 from kylemarsh/master
Sage Weil [Wed, 4 Jan 2012 23:53:51 +0000 (15:53 -0800)]
Merge pull request #8 from kylemarsh/master

Remove cloudfiles requirement from obsync.

13 years agoqa: load-gen-mix-small-long
Sage Weil [Wed, 4 Jan 2012 22:21:01 +0000 (14:21 -0800)]
qa: load-gen-mix-small-long

30 minutes

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoobsync: make obsync run without cloudfiles installed 8/head
Kyle Marsh [Wed, 4 Jan 2012 22:14:17 +0000 (14:14 -0800)]
obsync: make obsync run without cloudfiles installed

Cloudfiles probably shouldn't be a requirement for running obsync, so this
commit makes it optional.

13 years agoosd: fix misdirect check for requests with old epochs
Sage Weil [Wed, 4 Jan 2012 18:46:35 +0000 (10:46 -0800)]
osd: fix misdirect check for requests with old epochs

get_map() assumes the epoch passed is valid.  Check here in the caller.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosd: check that we're supposed to be getting a PG before waitlisting requests.
Sage Weil [Wed, 4 Jan 2012 18:44:12 +0000 (10:44 -0800)]
osd: check that we're supposed to be getting a PG before waitlisting requests.

This was broken in fa722de6708d3e92037df6289cc29ece12c8ea66.

Fix it by checking if the mapping was correct in the sender's epoch, and
either drop it (if valid) or handle_misdirected_request() if not.

Also fix the documentation for op_is_queueable() to not be a gigantic lie.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agorados: gracefully report errors from 'ls'
Sage Weil [Wed, 4 Jan 2012 18:40:06 +0000 (10:40 -0800)]
rados: gracefully report errors from 'ls'

Catch the exception thrown by the iterator when the OSD returns errors.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosd: return EINVAL on bad PGLS[_FILTER] handle
Sage Weil [Wed, 4 Jan 2012 17:49:28 +0000 (09:49 -0800)]
osd: return EINVAL on bad PGLS[_FILTER] handle

Fixes: #1875
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agolibrados: return int64_t pool ids
Josh Durgin [Wed, 4 Jan 2012 01:11:28 +0000 (17:11 -0800)]
librados: return int64_t pool ids

468e28ee60ee2fe625d2680c792a4bcb9ef19951 missed the get_id() functions.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agorados.py: use uint64_t for auids
Josh Durgin [Wed, 4 Jan 2012 00:24:59 +0000 (16:24 -0800)]
rados.py: use uint64_t for auids

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoradosgw-admin: add eol following info
Yehuda Sadeh [Tue, 3 Jan 2012 22:06:35 +0000 (14:06 -0800)]
radosgw-admin: add eol following info

13 years agotestrados: replace testreadwrite and testsnaps with testrados
Josh Durgin [Tue, 3 Jan 2012 19:09:00 +0000 (11:09 -0800)]
testrados: replace testreadwrite and testsnaps with testrados

testrados can act as testreadwrite or testsnaps by changing the
command line options for the weight of each operation type.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoRadosModel: check for out of order replies within WriteOps
Josh Durgin [Fri, 30 Dec 2011 02:36:54 +0000 (18:36 -0800)]
RadosModel: check for out of order replies within WriteOps

A single WriteOp already does multiple aio_writes. Each aio_write
gets a unique tid that is checked upon completion. There's no reason
to loop over the ranges twice since we can use the done flag instead
of the set of completions in WriteOp::finished().

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoRadosModel: allow TestOps to pass data to their finish methods
Josh Durgin [Fri, 30 Dec 2011 02:30:57 +0000 (18:30 -0800)]
RadosModel: allow TestOps to pass data to their finish methods

This will allow nested writes to keep track of which write actually
completed.  Also remove finish() and _finish() from TestOp subclasses
that had the same implementation as the superclass.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoRadosModel: make object write ranges configurable
Josh Durgin [Fri, 30 Dec 2011 01:55:45 +0000 (17:55 -0800)]
RadosModel: make object write ranges configurable

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoosd: add a monitor timeout via MPGStatsAck messages
Greg Farnum [Sat, 24 Dec 2011 00:41:38 +0000 (16:41 -0800)]
osd: add a monitor timeout via MPGStatsAck messages

Keep track of when we have outstanding updates, and while we do, make
sure the monitor responds within a timeout (default 30 seconds). If
it doesn't, reconnect!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agoFix invalid docdir_SCRIPTS usage with >=automake-1.11.2
Alphat-PC [Tue, 3 Jan 2012 15:36:15 +0000 (16:36 +0100)]
Fix invalid docdir_SCRIPTS usage with >=automake-1.11.2

13 years agolibrados: take lock in rollback
Sage Weil [Sat, 31 Dec 2011 01:04:47 +0000 (17:04 -0800)]
librados: take lock in rollback

We're poking through the osdmap; need to hold the lock here.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoobjecter: assert lock held in op_submit
Sage Weil [Sat, 31 Dec 2011 01:04:30 +0000 (17:04 -0800)]
objecter: assert lock held in op_submit

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agolibrados: call aio_operate() with lock held
Sage Weil [Sat, 31 Dec 2011 01:04:22 +0000 (17:04 -0800)]
librados: call aio_operate() with lock held

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agorgw: create default constructors for some structs
Yehuda Sadeh [Fri, 30 Dec 2011 22:18:40 +0000 (14:18 -0800)]
rgw: create default constructors for some structs

this will silence valgrind a bit

13 years agoworkunits: update rbd test for new error format
Josh Durgin [Fri, 30 Dec 2011 18:32:05 +0000 (10:32 -0800)]
workunits: update rbd test for new error format

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoconfig: use autoconf $libdir for default rados class dir
Sage Weil [Fri, 30 Dec 2011 17:50:36 +0000 (09:50 -0800)]
config: use autoconf $libdir for default rados class dir

Fixes: #1722
Signed-off-by: Sage Weil <sage@newdream.net>
13 years ago.gitignore: src/ocf/ceph
Sage Weil [Fri, 30 Dec 2011 17:17:06 +0000 (09:17 -0800)]
.gitignore: src/ocf/ceph

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoSpec: conditionally build ceph-resource-agents package
Florian Haas [Thu, 29 Dec 2011 19:58:02 +0000 (20:58 +0100)]
Spec: conditionally build ceph-resource-agents package

Put OCF resource agents in a separate subpackage,
to be enabled with a separate build conditional
(--with ocf).

Make the subpackage depend on the resource-agents
package, which provides the ocf-shellfuncs library
that the Ceph RAs use.

Signed-off-by: Florian Haas <florian@hastexo.com>
13 years agoAdd OCF-compliant resource agent for Ceph daemons
Florian Haas [Thu, 29 Dec 2011 19:58:01 +0000 (20:58 +0100)]
Add OCF-compliant resource agent for Ceph daemons

Add a wrapper around the ceph init script that makes
MDS, OSD and MON configurable as Open Cluster Framework
(OCF) compliant cluster resources. Allows Ceph
daemons to tie in with cluster resource managers that
support OCF, such as Pacemaker (http://www.clusterlabs.org).

Disabled by default, configure --with-ocf to enable.

Signed-off-by: Florian Haas <florian@hastexo.com>
13 years agomon: fix full ratio updates
Sage Weil [Fri, 30 Dec 2011 16:06:55 +0000 (08:06 -0800)]
mon: fix full ratio updates

- update them independently
- only if we are leader
- fix type for nearfull_ratio

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: don't ignore first full ratio update callback
Sage Weil [Fri, 30 Dec 2011 16:06:06 +0000 (08:06 -0800)]
mon: don't ignore first full ratio update callback

We get a callack on startup.  Don't ignore it.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: only update full_ratio if we're the leader
Sage Weil [Fri, 30 Dec 2011 15:45:21 +0000 (07:45 -0800)]
mon: only update full_ratio if we're the leader

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoMerge remote branch 'gh/wip-cleanup'
Sage Weil [Fri, 30 Dec 2011 15:42:20 +0000 (07:42 -0800)]
Merge remote branch 'gh/wip-cleanup'

13 years agomon: make full ratio config change callback safe
Sage Weil [Fri, 30 Dec 2011 01:15:07 +0000 (17:15 -0800)]
mon: make full ratio config change callback safe

We can't propose_pending() from any context; do this in the tick() thread,
with the proper locking.  Among other things, this fixes the crash on
startup that is now triggered due to eba235f2.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoclitests: update for new error format
Josh Durgin [Thu, 29 Dec 2011 23:43:55 +0000 (15:43 -0800)]
clitests: update for new error format

This was changed in 1f434da8a3ca4db830d1f3b0d87e5df941d85f2d

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoclitests: update monmaptool test
Josh Durgin [Thu, 29 Dec 2011 23:28:16 +0000 (15:28 -0800)]
clitests: update monmaptool test

e93961c11119942eae3a4cd14a79f779a5a4d277 changed output format.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>
13 years agoinit script: be LSB compliant for exit code on status
Florian Haas [Tue, 27 Dec 2011 10:43:47 +0000 (11:43 +0100)]
init script: be LSB compliant for exit code on status

An exit code of 1 on status is defined in LSB as
"program is dead, but pid file exists". Check for existence
of this pid file, and only set the exit status 1 if it's still there.
Set it to 3 ("program is not running") otherwise.

Reference: http://refspecs.linuxbase.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html

Signed-off-by: Florian Haas <florian@hastexo.com>
13 years agokeyring: print more useful errors to log/err
Sage Weil [Thu, 29 Dec 2011 19:58:28 +0000 (11:58 -0800)]
keyring: print more useful errors to log/err

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agocommon: trigger all observers on startup
Sage Weil [Thu, 29 Dec 2011 19:57:55 +0000 (11:57 -0800)]
common: trigger all observers on startup

Among other things, this makes err-to-stderr and friends initialize
properly in the DoutStreamBuf.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agocommon: make cpp_strerror output prettier
Sage Weil [Thu, 29 Dec 2011 19:24:50 +0000 (11:24 -0800)]
common: make cpp_strerror output prettier

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agolibrados: check for monclient::init() error
Sage Weil [Thu, 29 Dec 2011 19:24:12 +0000 (11:24 -0800)]
librados: check for monclient::init() error

I think this fixes #1835.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoqa: load-gen-mix-small.sh
Sage Weil [Thu, 29 Dec 2011 17:59:26 +0000 (09:59 -0800)]
qa: load-gen-mix-small.sh

13 years agorgw: turn on cache by default
Yehuda Sadeh [Thu, 29 Dec 2011 17:46:55 +0000 (09:46 -0800)]
rgw: turn on cache by default

13 years agomds: assert if we get an EINVAL on our truncate
Sage Weil [Wed, 30 Nov 2011 01:18:48 +0000 (17:18 -0800)]
mds: assert if we get an EINVAL on our truncate

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosd: get fsid from monmap, not osdmap
Sage Weil [Thu, 29 Dec 2011 17:00:46 +0000 (09:00 -0800)]
osd: get fsid from monmap, not osdmap

We may not have a valid OSDMap in all of these cases (notably, during
boot).  Always take the fsid from the monmap, which will be valid after
we've authenticated.

This fixes messages like

2011-12-29 08:53:44.530830 7ff3595e2700 mon.a@0(leader).pg v5 handle_statfs on fsid 00000000-0000-0000-0000-000000000000 != f8a6383d-5fbe-4f65-907e-f8d09e1d540d

on the monitor from MPGStats messages with a bad fsid right after osd boot.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomonc: get latest monmap during authentication
Sage Weil [Thu, 29 Dec 2011 16:59:00 +0000 (08:59 -0800)]
monc: get latest monmap during authentication

Tell the monitor which monmap version we have in our initial auth message.
Make the monitor send the latest monmap if it has something newer.  This
ensures that once authentication completes the monclient has the latest
monmap and a valid fsid.

Fixes: #1848
Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosd: catch authenticate error on startup
Sage Weil [Thu, 29 Dec 2011 16:43:41 +0000 (08:43 -0800)]
osd: catch authenticate error on startup

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoosdmap: add const markers to some unfixed functions
Greg Farnum [Fri, 23 Dec 2011 22:27:15 +0000 (14:27 -0800)]
osdmap: add const markers to some unfixed functions

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agorgw: removing swift user index when removing user
Yehuda Sadeh [Thu, 29 Dec 2011 01:00:19 +0000 (17:00 -0800)]
rgw: removing swift user index when removing user

13 years agorgw-admin: remove subuser index when required
Yehuda Sadeh [Thu, 29 Dec 2011 00:50:36 +0000 (16:50 -0800)]
rgw-admin: remove subuser index when required

13 years agorgw: clone operation should only update index for main category
Yehuda Sadeh [Thu, 29 Dec 2011 00:34:03 +0000 (16:34 -0800)]
rgw: clone operation should only update index for main category

13 years agorgw: fix cache interface (was not overloading method)
Yehuda Sadeh [Thu, 29 Dec 2011 00:33:14 +0000 (16:33 -0800)]
rgw: fix cache interface (was not overloading method)

13 years agorgw: fix bucket creation
Yehuda Sadeh [Wed, 28 Dec 2011 23:10:10 +0000 (15:10 -0800)]
rgw: fix bucket creation

13 years agorgw: write bucket info in one operation
Yehuda Sadeh [Fri, 23 Dec 2011 20:46:05 +0000 (12:46 -0800)]
rgw: write bucket info in one operation

13 years agoObjecter: fix local reads one more time.
Greg Farnum [Fri, 23 Dec 2011 17:56:42 +0000 (09:56 -0800)]
Objecter: fix local reads one more time.

Document it a little since we've gotten it wrong so often.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>
13 years agofilestore: fix config observer
Sage Weil [Thu, 22 Dec 2011 23:25:00 +0000 (15:25 -0800)]
filestore: fix config observer

Actually, I don't think this was fully implemented to begin with, so it's
not a 'fix' per se.  This will let you use injectargs to adjust the
filestore config options during runtime.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoMOSDPGRepScrub: Fix typo in MOSDPGRepScrub
Samuel Just [Thu, 22 Dec 2011 23:17:22 +0000 (15:17 -0800)]
MOSDPGRepScrub: Fix typo in MOSDPGRepScrub

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
13 years agoMerge branch 'wip-signal'
Sage Weil [Thu, 22 Dec 2011 22:00:23 +0000 (14:00 -0800)]
Merge branch 'wip-signal'

13 years agoMOSDRepScrub: use header.version for payload version
Samuel Just [Thu, 22 Dec 2011 20:33:48 +0000 (12:33 -0800)]
MOSDRepScrub: use header.version for payload version

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
13 years agoMerge branch 'stable'
Sage Weil [Thu, 22 Dec 2011 20:16:06 +0000 (12:16 -0800)]
Merge branch 'stable'

13 years agomonmap: iterate over addr_name when printing summary
Henry C Chang [Thu, 22 Dec 2011 10:48:29 +0000 (18:48 +0800)]
monmap: iterate over addr_name when printing summary

The rank is now ordered by IP address. We should iterate over
addr_name.

Signed-off-by: Henry C Chang <henry.cy.chang@gmail.com>
13 years agomonmap: clear addr_name map on calculating ranks
Henry C Chang [Thu, 22 Dec 2011 10:49:04 +0000 (18:49 +0800)]
monmap: clear addr_name map on calculating ranks

We should clear addr_name before filling it. Otherwise, the removed
mon will stay there and cause incorrect rank assignment.

Signed-off-by: Henry C Chang <henry.cy.chang@gmail.com>
13 years agointerval_set: fix truncation of _size
Henry C Chang [Thu, 22 Dec 2011 10:47:25 +0000 (18:47 +0800)]
interval_set: fix truncation of _size

_size is type of int64_t. Use int to store the value of _size
will cause value truncation.

Signed-off-by: Henry C Chang <henry.cy.chang@gmail.com>
13 years agoosd: remove SIGTERM cruft
Sage Weil [Wed, 21 Dec 2011 22:21:37 +0000 (14:21 -0800)]
osd: remove SIGTERM cruft

The default handler will exit(0).  The got_sigterm stuff was dead code.

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomon: drop special SIGTERM handler
Sage Weil [Wed, 21 Dec 2011 22:17:42 +0000 (14:17 -0800)]
mon: drop special SIGTERM handler

Default does exit(0).

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agomds: drop special SIGTERM handler
Sage Weil [Wed, 21 Dec 2011 22:17:19 +0000 (14:17 -0800)]
mds: drop special SIGTERM handler

Default does exit(0).

Signed-off-by: Sage Weil <sage@newdream.net>
13 years agoexit(0) on SIGTERM by default
Sage Weil [Wed, 21 Dec 2011 22:17:00 +0000 (14:17 -0800)]
exit(0) on SIGTERM by default

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