Commit Graph

25155 Commits

Author SHA1 Message Date
Yehuda Sadeh
b07f3cda5e rgw: add encode_json
dump() just dumps the internal content of an object, encode_json()
create the object inside its own section. Note that there are cases
where we don't want an object to be surrounded by a section, e.g.,
when an object represents an array. In such a case we'd need to
override the encode_json() for this object type.

Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2013-03-22 11:23:08 -07:00
Yehuda Sadeh
96ad9b149f moved rgw/rgw_json.* common/ceph_json.*
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2013-03-22 11:23:08 -07:00
Yehuda Sadeh
53ba6c2d0f rgw: rgw_jsonparser changes
Now testing json_decode functionality.

Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2013-03-22 11:23:07 -07:00
Yehuda Sadeh
da33701318 rgw: decode RGWUserInfo using decode_json
RGWUserInfo can be decoded using json_decode, and fix
the ->dump() method to match that decoding.

Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2013-03-22 11:23:07 -07:00
Yehuda Sadeh
959b658fd0 rgw: introduce decode_json to easily decode json structures
Introducing a new way to define the json structure of an
object, similar to what we do with bufferlist encode/decode.

Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
2013-03-22 11:23:07 -07:00
Sage Weil
6d63752c8f ceph-disk: re-add python 2.7 dependency comment
FIXME!

Signed-off-by: Sage Weil <sage@inktank.com>
2013-03-22 10:09:55 -07:00
Sage Weil
c9157a1946 Merge pull request #117 from ceph/wip-ceph-disk
ceph-disk-* refactor
2013-03-22 10:06:13 -07:00
Sage Weil
0981e4666b Merge branch 'next' 2013-03-22 09:15:52 -07:00
Sage Weil
38a5acbb82 osd: reenable 'journal aio = true'
Now that #4079 is resolved.  Reverts 1cfc3ae0.

Signed-off-by: Sage Weil <sage@inktank.com>
2013-03-22 09:15:23 -07:00
Sage Weil
e5940da9a5 os/FileJournal: fix aio self-throttling deadlock
This block of code tries to limit the number of aios in flight by waiting
for the amount of data to be written to grow relative to a function of the
number of aios.  Strictly speaking, the condition we are waiting for is a
function of both aio_num and the write queue, but we are only woken by
changes in aio_num, and were (in rare cases) waiting when aio_num == 0 and
there was no possibility of being woken.

Fix this by verifying that aio_num > 0, and restructuring the loop to
recheck that condition on each wakeup.

Fixes: #4079
Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Samuel Just <sam.just@inktank.com>
2013-03-22 09:15:20 -07:00
Sage Weil
a35b865093 Merge pull request #137 from dalgaaf/wip-da-cleanup-includes
Cleanup some twice included header

Reviewed-by: Sage Weil <sage@inktank.com>
2013-03-22 08:46:31 -07:00
Danny Al-Gaaf
6c79604cef test/test_snap_mapper.cc: remove twice included <tr1/memory>
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
2013-03-22 16:03:22 +01:00
Danny Al-Gaaf
00cf8178b6 mon/MDSMonitor.cc: remove twice included MonitorDBStore.h
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
2013-03-22 16:02:55 +01:00
Danny Al-Gaaf
a583029e81 mon/LogMonitor.cc: remove twice included <sstream>
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
2013-03-22 16:02:23 +01:00
Danny Al-Gaaf
9dd5b209a2 mon/AuthMonitor.cc: remove twice included <sstream>
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
2013-03-22 16:01:53 +01:00
Danny Al-Gaaf
1144260957 common/Formatter.h: remove twice included <list>
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
2013-03-22 16:01:15 +01:00
Samuel Just
000310fd10 ReplicatedPG: add debug flag to skip full check at reservation
This will make it easier to test the check in do_scan.

Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:40 -07:00
Samuel Just
29a288f993 ReplicatedPG: replica should post BackfillTooFull in do_scan if full
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:40 -07:00
Samuel Just
f9c8190e4f PG: halt backfill on RemoteReservationRejected in Backilling
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:40 -07:00
Samuel Just
022903bf73 PG: add helper for adding a timer event to retry backfill
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:40 -07:00
Samuel Just
3ab4a09de2 PG: add BackfillTooFull event for RepRecovering
Replica will use this to notify Primary to stop backfilling.

Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:39 -07:00
Samuel Just
cd7353938e PG: add helper for rejecting backfill reservation
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:39 -07:00
Samuel Just
33aaea594d PG: use OSDService::too_full_for_backfill in RepWaitBackfillReserved
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:39 -07:00
Samuel Just
cec3d82ec4 OSDService: add too_full_for_backfill
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:51:39 -07:00
Samuel Just
f00f3bc4e5 Merge remote-tracking branch 'upstream/wip_osd_shutdown_notification'
Fixes: #1857
Fixes: #4267
Reviewed-by: Sage Weil <sage@inktank.com>
2013-03-21 18:46:50 -07:00
Samuel Just
fab0be1ffe Makefile: add MOSDMarkMeDown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:46:14 -07:00
Samuel Just
7bfaaf7a20 OSD: notify mon prior to shutdown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:46:11 -07:00
Samuel Just
a2dba959d2 Monitor: add MOSDMarkMeDown support
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
edebb341c2 OSDMonitor: factor out check_source helper
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
b8929c4262 messages: add MOSDMarkMeDown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
bb125221be OSD: clear OpHistory on shutdown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
51ae509800 OpRequest: use OpRequestRef for OpHistory
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
210ecc5871 FileStore::stat: valgrind: don't read *st on error
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
6119f297dc shared_cache: clear lru in destructor
Otherwise, the live references will attempt to extricate
themselves from a disolving SharedLRU instance as the
member destructors run.

Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
95e9322c65 ceph_osd: clear client_throttler prior to putting g_ceph_context
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
8d43a2ec91 OSD: reorder OSD::shutdown
Reorder teardown:
- pgs
- queues/threadpools
- persist superblock
- filestore
- timers
- messengers

Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:35 -07:00
Samuel Just
505bebbb3c OSD: check for is_stopping after locking osd_lock or heartbeat_lock
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Samuel Just
9d076a0edb OSD: lookup_lock_raw_pg is dead
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Samuel Just
e41554da14 OSD: rename timer to tick_timer
Only used for scheduling ticks - we should keep it
that way.

Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Samuel Just
30b7a151f3 OSD: move backfill_request_timer cleanup to OSDService::shutdown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Samuel Just
264cd3efca ReplicatedPG,OSD: move dequeueing into on_shutdown
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Samuel Just
eb8dd0d404 PG: clarify PG::deleting comment
Signed-off-by: Samuel Just <sam.just@inktank.com>
2013-03-21 18:37:34 -07:00
Sage Weil
7118df89cd Merge pull request #135 from ceph/wip-4519
mon: AuthMonitor: delete auth_handler while increasing max_global_id

Reviewed-by: Sage Weil <sage@inktank.com>
2013-03-21 18:25:01 -07:00
Joao Eduardo Luis
71ec9c6bd5 mon: AuthMonitor: delete auth_handler while increasing max_global_id
By not deleting and setting NULL the session's auth_handler, we could
hit a scenario in which we'd end up dispatching a previously-wait-listed
auth message and we wouldn't start its auth session.

This only happened when increasing max_global_id via Paxos (in which case
we would wait-list the message) and would only be noticeable when running
with cephx disabled.

Fixes: #4519

Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
2013-03-22 01:21:00 +00:00
Samuel Just
42a71c1dd8 FileJournal: quieter debugging on journal scanning
Signed-off-by: Samuel Just <sam.just@inktank.com>
Reviewed-by: David Zafman <david.zafman@inktank.com>
(cherry picked from commit 6740d512ac)
2013-03-21 18:09:58 -07:00
Samuel Just
6740d512ac FileJournal: quieter debugging on journal scanning
Signed-off-by: Samuel Just <sam.just@inktank.com>
Reviewed-by: David Zafman <david.zafman@inktank.com>
2013-03-21 18:03:14 -07:00
John Wilkins
7e8cc57dee doc: Fixed some typos.
Signed-off-by: John Wilkins <john.wilkins@inktank.com>
2013-03-21 17:04:54 -07:00
Josh Durgin
56183ebc86 Merge pull request #128 from dalgaaf/wip-da-pylint
Fix some python issues

Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
2013-03-21 16:41:34 -07:00
Sage Weil
7a4b880f5d Merge pull request #132 from ceph/wip-libcephfs-fix
libcephfs: fix ceph_get_osd_crush_location

Reviewed-by: Sage Weil <sage@inktank.com>
2013-03-21 16:16:31 -07:00
Noah Watkins
76985007fd libcephfs: fix ceph_get_osd_crush_location
(a) Allow NULL buffer when length is zero to get the target buffer
length. (b) fix edge case where buffer length is the exact size
required.

Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
2013-03-21 15:32:39 -07:00