commit d783e33b672ec324eb48d588f956da0c51ff5dac Author: Gary Lowell Date: Sun Aug 18 23:54:49 2013 -0700 v0.61.8 commit 21a6e2479133a3debb9ab9057ff9fae70c9eede9 Author: Samuel Just Date: Thu Aug 8 15:12:46 2013 -0700 RadosClient: shutdown monclient after dropping lock Otherwise, the monclient shutdown may deadlock waiting on a context trying to take the RadosClient lock. Fixes: #5897 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 0aacd10e2557c55021b5be72ddf39b9cea916be4) commit 64bef4ae4bab28b0b82a1481381b0c68a22fe1a4 Author: Sage Weil Date: Sat Aug 17 09:05:32 2013 -0700 mon/OSDMonitor: make 'osd pool mksnap ...' not expose uncommitted state [This is a backport of d1501938f5d07c067d908501fc5cfe3c857d7281] We were returning success without waiting if the pending pool state had the snap. Signed-off-by: Sage Weil Reviewed-by: Joao Eduardo Luis commit 411871f6bcc9a4b81140c2e98d13dc123860f6f7 Author: Sage Weil Date: Fri Aug 16 10:52:02 2013 -0700 mon/OSDMonitor: make 'osd pool rmsnap ...' not racy/crashy NOTE: This is a manual backport of d90683fdeda15b726dcf0a7cab7006c31e99f14. Due to all kinds of collateral changes in the mon the original patch doesn't cleanly apply. Ensure that the snap does in fact exist before we try to remove it. This avoids a crash where a we get two dup rmsnap requests (due to thrashing, or a reconnect, or something), the committed (p) value does have the snap, but the uncommitted (pp) does not. This fails the old test such that we try to remove it from pp again, and assert. Restructure the flow so that it is easier to distinguish the committed short return from the uncommitted return (which must still wait for the commit). 0> 2013-07-16 14:21:27.189060 7fdf301e9700 -1 osd/osd_types.cc: In function 'void pg_pool_t::remove_snap(snapid_t)' thread 7fdf301e9700 time 2013-07-16 14:21:27.187095 osd/osd_types.cc: 662: FAILED assert(snaps.count(s)) ceph version 0.66-602-gcd39d8a (cd39d8a6727d81b889869e98f5869e4227b50720) 1: (pg_pool_t::remove_snap(snapid_t)+0x6d) [0x7ad6dd] 2: (OSDMonitor::prepare_command(MMonCommand*)+0x6407) [0x5c1517] 3: (OSDMonitor::prepare_update(PaxosServiceMessage*)+0x1fb) [0x5c41ab] 4: (PaxosService::dispatch(PaxosServiceMessage*)+0x937) [0x598c87] 5: (Monitor::handle_command(MMonCommand*)+0xe56) [0x56ec36] 6: (Monitor::_ms_dispatch(Message*)+0xd1d) [0x5719ad] 7: (Monitor::handle_forward(MForward*)+0x821) [0x572831] 8: (Monitor::_ms_dispatch(Message*)+0xe44) [0x571ad4] 9: (Monitor::ms_dispatch(Message*)+0x32) [0x588c52] 10: (DispatchQueue::entry()+0x549) [0x7cf1d9] 11: (DispatchQueue::DispatchThread::entry()+0xd) [0x7060fd] 12: (()+0x7e9a) [0x7fdf35165e9a] 13: (clone()+0x6d) [0x7fdf334fcccd] NOTE: a copy of the executable, or `objdump -rdS ` is needed to interpret this. Signed-off-by: Sage Weil Reviewed-by: Joao Eduardo Luis commit 50698d1862065c8d74338fd08c7e5af66e222490 Author: Sage Weil Date: Tue Aug 13 12:52:41 2013 -0700 librados: fix async aio completion wakeup For aio flush, we register a wait on the most recent write. The write completion code, however, was *only* waking the waiter if they were waiting on that write, without regard to previous writes (completed or not). For example, we might have 6 and 7 outstanding and wait on 7. If they finish in order all is well, but if 7 finishes first we do the flush completion early. Similarly, if we - start 6 - start 7 - finish 7 - flush; wait on 7 - finish 6 we can hang forever. Fix by doing any completions that are prior to the oldest pending write in the aio write completion handler. Refs: #5919 Signed-off-by: Sage Weil Reviewed-by: Josh Durgin Tested-by: Oliver Francke (cherry picked from commit 16ed0b9af8bc08c7dabead1c1a7c1a22b1fb02fb) commit ef731dfc84a71d3c3262f5cff9a9d33a60255485 Author: Josh Durgin Date: Mon Aug 12 19:17:09 2013 -0700 librados: fix locking for AioCompletionImpl refcounting Add an already-locked helper so that C_Aio{Safe,Complete} can increment the reference count when their caller holds the lock. C_AioCompleteAndSafe's caller is not holding the lock, so call regular get() to ensure no racing updates can occur. This eliminates all direct manipulations of AioCompletionImpl->ref, and makes the necessary locking clear. The only place C_AioCompleteAndSafe is used is in handling aio_flush_async(). This could cause a missing completion. Refs: #5919 Signed-off-by: Josh Durgin Reviewed-by: Sage Weil Tested-by: Oliver Francke (cherry picked from commit 7a52e2ff5025754f3040eff3fc52d4893cafc389) commit 32631685199f2e47c2ba0ed27d16eff80fa6917d Author: Sage Weil Date: Fri Jul 12 14:27:04 2013 -0700 mon/Paxos: bootstrap peon too if monmap updates If we get a monmap update, the leader bootstraps. Peons should do the same. Signed-off-by: Sage Weil Reviewed-by: Greg Farnum (cherry picked from commit efe5b67bb700ef6218d9579abf43cc9ecf25ef52) commit 1ea6b56170fc9e223e7c30635db02fa2ad8f4b4e Author: Sage Weil Date: Tue Jun 25 13:16:45 2013 -0700 osd: fix race when queuing recovery ops Previously we would sample how many ops to start under the lock, drop it, and start that many. This is racy because multiple threads can jump in and we start too many ops. Instead, claim as many slots as we can and release them back later if we do not end up using them. Take care to re-wake the work-queue since we are releasing more resources for wq use. Signed-off-by: Sage Weil Reviewed-by: Samuel Just (cherry picked from commit 01d3e094823d716be0b39e15323c2506c6f0cc3b) commit 4433f9ad8b338b6a55e205602434b307287bfaa3 Author: Sage Weil Date: Mon Jun 24 16:37:29 2013 -0700 osd: tolerate racing threads starting recovery ops We sample the (max - active) recovery ops to know how many to start, but do not hold the lock over the full duration, such that it is possible to start too many ops. This isn't problematic except that our condition checks for being == max but not beyond it, and we will continue to start recovery ops when we shouldn't. Fix this by adjusting the conditional to be <=. Reported-by: Stefan Priebe Signed-off-by: Sage Weil Reviewed-by: David Zafman (cherry picked from commit 3791a1e55828ba541f9d3e8e3df0da8e79c375f9) commit 0964d53ef3e8e386e0a1635d2240aefad7b8e2c1 Author: Sage Weil Date: Fri Aug 9 18:02:32 2013 -0700 ceph-disk: fix mount options passed to move_mount Commit 6cbe0f021f62b3ebd5f68fcc01a12fde6f08cff5 added a mount_options but in certain cases it may be blank. Fill in with the defaults, just as we do in mount(). Backport: cuttlefish Reviewed-by: Dan Mick Signed-off-by: Sage Weil (cherry picked from commit cb50b5a7f1ab2d4e7fdad623a0e7769000755a70) commit d6be5ed2601b8cf45570afe7ca75ce5aba3f8b4f Author: Yehuda Sadeh Date: Mon Aug 12 10:05:44 2013 -0700 rgw: fix multi delete Fixes: #5931 Backport: bobtail, cuttlefish Fix a bad check, where we compare the wrong field. Instead of comparing the ret code to 0, we compare the string value to 0 which generates implicit casting, hence the crash. Signed-off-by: Yehuda Sadeh Reviewed-by: Josh Durgin (cherry picked from commit f9f1c48ad799da2b4be0077bf9d61ae116da33d7) Conflicts: src/rgw/rgw_rest_s3.cc commit ecaa46a13837305b9382ab319d43890729c54f1e Author: Danny Al-Gaaf Date: Tue Jul 23 21:56:09 2013 +0200 ceph.spec.in: obsolete ceph-libs only on the affected distro The ceph-libs package existed only on Redhat based distro, there was e.g. never such a package on SUSE. Therefore: make sure the 'Obsoletes' is only set on these affected distros. Signed-off-by: Danny Al-Gaaf commit 81aa68c309a6d4eaecc54f8d735efde8843fed8c Author: Gary Lowell Date: Wed Jul 3 11:28:28 2013 -0700 ceph.spec.in: Obsolete ceph-libs Signed-off-by: Gary Lowell Reviewed-by: Sage Weil commit 2a34df68bb02d14f6a25bd13dff600a4d629ad05 Author: Joao Eduardo Luis Date: Fri Aug 9 14:48:15 2013 -0700 common: pick_addresses: fix bug with observer class that triggered #5205 The Observer class we defined to observe conf changes and thus avoid triggering #5205 (as fixed by eb86eebe1ba42f04b46f7c3e3419b83eb6fe7f9a), was returning always the same const static array, which would lead us to always populate the observer's list with an observer for 'public_addr'. This would of course become a problem when trying to obtain the observer for 'cluster_add' during md_config_t::set_val() -- thus triggering the same assert as initially reported on #5205. Backport: cuttlefish Fixes: #5205 Signed-off-by: Joao Eduardo Luis Reviewed-by: Sage Weil (cherry picked from commit 7ed6de9dd7aed59f3c5dd93e012cf080bcc36d8a) commit 1243c9749ed27850c5d041023780efcdf7b31a68 Author: Alfredo Deza Date: Thu Aug 8 16:09:26 2013 -0700 make sure we are using the mount options Signed-off-by: Alfredo Deza (cherry picked from commit 34831d0989d4bcec4920068b6ee09ab6b3234c91) commit a9a370be2d8155b696ebe2866febb0571da5740f Author: Samuel Just Date: Fri Aug 2 11:58:52 2013 -0700 PG: set !flushed in Reset() Otherwise, we might serve a pull before we start_flush in the ReplicaActive constructor. Fixes: #5799 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 9e7d6d547e0e8a6db6ba611882afa9bf74ea0195) commit 65bfa4941f983c988837cd010f731966ff53fd19 Author: Sage Weil Date: Fri Jul 26 14:02:07 2013 -0700 osd: make open classes on start optional This is cuttlefish; default to the old behavior! Signed-off-by: Sage Weil (cherry picked from commit 6f996223fb34650771772b88355046746f238cf2) commit e8253ae5451b1c8e3d7d50199b8db7b2d4c66486 Author: Sage Weil Date: Fri Jul 26 13:58:46 2013 -0700 osd: load all classes on startup This avoid creating a wide window between when ceph-osd is started and when a request arrives needing a class and it is loaded. In particular, upgrading the packages in that window may cause linkage errors (if the class API has changed, for example). Fixes: #5752 Signed-off-by: Sage Weil Reviewed-by: Yehuda Sadeh (cherry picked from commit c24e652d8c5e693498814ebe38c6adbec079ea36) Conflicts: src/osd/ClassHandler.cc commit 7a1d6d3e727fd8b6947c658e171bf7ec31cd7966 Author: Sage Weil Date: Sun Jul 28 15:42:08 2013 -0700 ceph_test_rados: print version banner on startup It is helpful when looking at qa run logs to see what version of the tester is running. Signed-off-by: Sage Weil (cherry picked from commit 12c1f1157c7b9513a3d9f716a8ec62fce00d28f5) commit 86769f05ccc54bfec403bb9ea9a3a951bbcea301 Author: Sage Weil Date: Thu Jun 13 22:08:36 2013 -0700 ceph_test_rados: add --pool arg Signed-off-by: Sage Weil (cherry picked from commit bcfbd0a3ffae6947464d930f636c8b35d1331e9d) commit b70a9abc5e3ae01204256f414bd7e69d083ed7c6 Author: Sage Weil Date: Fri Jul 26 14:07:02 2013 -0700 upstart: stop ceph-create-keys when the monitor stops This avoids lingering ceph-create-keys tasks. Backport: cuttlefish Signed-off-by: Sage Weil (cherry picked from commit a90a2b42db8de134b8ea5d81cab7825fb9ec50b4) commit 5af48dc7c7e3a0d7f7bc22af58831d58d165e657 Author: Samuel Just Date: Fri Jul 26 13:42:27 2013 -0700 FileStore: fix fd leak in _check_global_replay_guard Bug introduced in f3f92fe21061e21c8b259df5ef283a61782a44db. Fixes: #5766 Backport: cuttlefish Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit c562b72e703f671127d0ea2173f6a6907c825cd1) commit 17aa2d6d16c77028bae1d2a77903cdfd81efa096 Author: Sage Weil Date: Thu Jul 25 11:10:53 2013 -0700 mon/Paxos: share uncommitted value when leader is/was behind If the leader has and older lc than we do, and we are sharing states to bring them up to date, we still want to also share our uncommitted value. This particular case was broken by b26b7f6e, which was only contemplating the case where the leader was ahead of us or at the same point as us, but not the case where the leader was behind. Note that the call to share_state() a few lines up will bring them fully up to date, so after they receive and store_state() for this message they will be at the same lc as we are. Fixes: #5750 Backport: cuttlefish Signed-off-by: Sage Weil Reviewed-by: Greg Farnum Reviewed-by: Joao Eduardo Luis (cherry picked from commit 05b6c7e8645081f405c616735238ae89602d3cc6) commit 09a664e25391dbad9a479bae33904d28231f429d Merge: 8f010af b0535fc Author: Sage Weil Date: Thu Jul 25 15:21:31 2013 -0700 Merge remote-tracking branch 'gh/cuttlefish-next' into cuttlefish commit b0535fcf854c5042d6b5ff481aabca08026d8f7f Author: Samuel Just Date: Tue Jul 23 18:04:40 2013 -0700 HashIndex: reset attr upon split or merge completion A replay of an in progress merge or split might make our counts unreliable. Fixes: #5723 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 0dc3efdd885377a07987d868af5bb7a38245c90b) commit 8f73302b4e637ca8b85d68ea7503279faecb57d8 Author: Samuel Just Date: Tue Jul 23 17:34:25 2013 -0700 test/filestore/store_test: add test for 5723 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 37a4c4af54879512429bb114285bcb4c7c3488d5) Conflicts: src/os/LFNIndex.cc src/test/filestore/store_test.cc commit 6a7b9e5f0c1d2344209c69ab9992f94221a16468 Author: Samuel Just Date: Tue Jul 23 13:51:26 2013 -0700 FileStore::_collection_rename: fix global replay guard If the replay is being replayed, we might have already performed the rename, skip it. Also, we must set the collection replay guard only after we have done the rename. Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 870c474c5348831fcb13797d164f49682918fb30) commit 7d98651775265896c22bacfc4afcfccbb0128470 Author: Samuel Just Date: Mon Jul 22 13:46:10 2013 -0700 PGLog::rewind_divergent_log: unindex only works from tail, index() instead Fixes: #5714 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 6957dbc75cc2577652b542aa3eae69f03060cb63) The original patch covered the same code in PGLog.cc. Conflicts: src/osd/PGLog.cc src/osd/PG.cc commit 611a06ae6c9cba468db206dfc82ec883c7a394af Author: Sage Weil Date: Thu Jul 18 09:55:43 2013 -0700 msg/Pipe: do not hold pipe_lock for verify_authorizer() We shouldn't hold the pipe_lock while doing the ms_verify_authorizer upcalls. Fix by unlocking a bit earlier, and verifying our state is still correct in the failure path. This regression was introduced by ecab4bb9513385bd765cca23e4e2fadb7ac4bac2. Signed-off-by: Sage Weil Reviewed-by: Greg Farnum (cherry picked from commit 723d691f7a1f53888618dfc311868d1988f61f56) Conflicts: src/msg/Pipe.cc commit 45bda482fa8a23f4b80d115e29d6f04cb5e226d6 Author: Sage Weil Date: Tue Jul 16 14:17:05 2013 -0700 msg/Pipe: a bit of additional debug output Signed-off-by: Sage Weil (cherry picked from commit 16568d9e1fb8ac0c06ebaa1e1dc1d6a432a5e4d4) commit 806eab59ad1a32aedb662c51de3b4a1d61fcbb62 Author: Sage Weil Date: Tue Jul 16 13:13:46 2013 -0700 msg/Pipe: hold pipe_lock during important parts of accept() Previously we did not bother with locking for accept() because we were not visible to any other threads. However, we need to close accepting Pipes from mark_down_all(), which means we need to handle interference. Fix up the locking so that we hold pipe_lock when looking at Pipe state and verify that we are still in the ACCEPTING state any time we retake the lock. Signed-off-by: Sage Weil (cherry picked from commit ecab4bb9513385bd765cca23e4e2fadb7ac4bac2) commit ce6a0b74459996f91a0511a4a7147179bcd47876 Author: Greg Farnum Date: Wed Jul 17 15:23:12 2013 -0700 msgr: fix a typo/goto-cross from dd4addef2d We didn't build or review carefully enough! Signed-off-by: Greg Farnum Reviewed-by: Sage Weil (cherry picked from commit 1a84411209b13084b3edb87897d5d678937e3299) commit 1ed51ad535612d5c444a3cc35a331f5e6a68ce30 Author: Sage Weil Date: Mon Jul 15 17:16:23 2013 -0700 msgr: close accepting_pipes from mark_down_all() We need to catch these pipes too, particularly when doing a rebind(), to avoid them leaking through. Signed-off-by: Sage Weil (cherry picked from commit 687fe888b32ac9d41595348dfc82111c8dbf2fcb) commit 2f696f17a413015a3038d5aa76d18fe94f503f03 Author: Sage Weil Date: Mon Jul 15 17:14:25 2013 -0700 msgr: maintain list of accepting pipes New pipes exist in a sort of limbo before we know who the peer is and add them to rank_pipe. Keep a list of them in accepting_pipes for that period. Signed-off-by: Sage Weil (cherry picked from commit dd4addef2d5b457cc9a58782fe42af6b13c68b81) commit 540a6f49d402c1990f0e0fe9f8897dd664e79501 Author: Sage Weil Date: Tue Jul 16 16:25:28 2013 -0700 msgr: adjust nonce on rebind() We can have a situation where: - we have a pipe to a peer - pipe goes to standby (on peer) - we rebind to a new port - .... - we rebind again to the same old port - we connect to peer and get reattached to the ancient pipe from two instances back. Avoid that by picking a new nonce each time we rebind. Add 1,000,000 each time so that the port is still legible in the printed output. Signed-off-by: Sage Weil (cherry picked from commit 994e2bf224ab7b7d5b832485ee14de05354d2ddf) Conflicts: src/msg/Accepter.cc commit f938a5bf604885ffba65a9b86e19258ca254e58c Author: Sage Weil Date: Mon Jul 15 17:10:23 2013 -0700 msgr: mark_down_all() after, not before, rebind If we are shutting down all old connections and binding to new ports, we want to avoid a sequence like: - close all prevoius connections - new connection comes in on old port - rebind to new ports -> connection from old port leaks through As a first step, close all connections after we shut down the old accepter and before we start the new one. Signed-off-by: Sage Weil (cherry picked from commit 07a0860a1899c7353bb506e33de72fdd22b857dd) Conflicts: src/msg/SimpleMessenger.cc commit 07b9ebf4212d53606ce332ff927a2ff68ed26978 Author: Sage Weil Date: Tue Jul 16 13:01:18 2013 -0700 msg/Pipe: unlock msgr->lock earlier in accept() Small cleanup. Nothing needs msgr->lock for the previously larger window. Signed-off-by: Sage Weil (cherry picked from commit ad548e72fd94b4a16717abd3b3f1d1be4a3476cf) commit ae85a0a101d624363fe761c06ecd52d3d38ba4a2 Author: Sage Weil Date: Tue Jul 16 10:09:02 2013 -0700 msg/Pipe: avoid creating empty out_q entry We need to maintain the invariant that all sub queues in out_q are never empty. Fix discard_requeued_up_to() to avoid creating an entry unless we know it is already present. This bug leads to an incorrect reconnect attempt when - we accept a pipe (lossless peer) - they send some stuff, maybe - fault - we initiate reconnect, even tho we have nothing queued In particular, we shouldn't reconnect because we aren't checking for resets, and the fact that our out_seq is 0 while the peer's might be something else entirely will trigger asserts later. This fixes at least one source of #5626, and possibly #5517. Backport: cuttlefish Signed-off-by: Sage Weil (cherry picked from commit 9f1c27261811733f40acf759a72958c3689c8516) commit 21e27262edc6f5f090ea8915517ee867e30b9066 Author: Sage Weil Date: Mon Jul 15 14:47:05 2013 -0700 msg/Pipe: assert lock is held in various helpers These all require that we hold pipe_lock. Signed-off-by: Sage Weil (cherry picked from commit 579d858aabbe5df88543d096ef4dbddcfc023cca) commit 25f4786ac41869b3f135bd072000634765bb8919 Author: Sage Weil Date: Sun Jul 14 08:55:52 2013 -0700 msg/Pipe: be a bit more explicit about encoding outgoing messages Signed-off-by: Sage Weil (cherry picked from commit 4282971d47b90484e681ff1a71ae29569dbd1d32) commit 48105a32605aa59b6970eb89fce4ecc4201e8d04 Author: Sage Weil Date: Fri Jul 12 16:21:24 2013 -0700 msg/Pipe: fix RECONNECT_SEQ behavior Calling handle_ack() here has no effect because we have already spliced sent messages back into our out queue. Instead, pull them out of there and discard. Add a few assertions along the way. Signed-off-by: Sage Weil Reviewed-by: Greg Farnum (cherry picked from commit 495ee108dbb39d63e44cd3d4938a6ec7d11b12e3) commit 1eab069017ce6b71e4bc2bb9679dbe31b50ae938 Author: Sage Weil Date: Mon Jun 17 13:32:38 2013 -0700 msgr: reaper: make sure pipe has been cleared (under pipe_lock) All paths to pipe shutdown should have cleared the con->pipe reference already. Assert as much. Also, do it under pipe_lock! Signed-off-by: Sage Weil (cherry picked from commit 9586305a2317c7d6bbf31c9cf5b67dc93ccab50d) commit db06a5092bc45d0479fe492a5d592713a7c53494 Author: Sage Weil Date: Mon Jun 17 14:14:02 2013 -0700 msg/Pipe: goto fail_unlocked on early failures in accept() Instead of duplicating an incomplete cleanup sequence (that does not clear_pipe()), goto fail_unlocked and do the cleanup in a generic way. s/rc/r/ while we are here. Signed-off-by: Sage Weil (cherry picked from commit ec612a5bda119cea52bbac9b2a49ecf1e83b08e5) commit 8612e50fd70bfceebd6c291e6cab10d9dfd39e8c Author: Sage Weil Date: Mon Jun 17 13:32:07 2013 -0700 msgr: clear con->pipe inside pipe_lock on mark_down We need to do this under protection of the pipe_lock. Signed-off-by: Sage Weil (cherry picked from commit afafb87e8402242d3897069f4b94ba46ffe0c413) commit 8aafe131acadc22cb069f3d98bba6922ab09c749 Author: Sage Weil Date: Mon Jun 17 12:47:11 2013 -0700 msgr: clear_pipe inside pipe_lock on mark_down_all Observed a segfault in rebind -> mark_down_all -> clear_pipe -> put that may have been due to a racing thread clearing the connection_state pointer. Do the clear_pipe() call under the protection of pipe_lock, as we do in all other contexts. Signed-off-by: Sage Weil (cherry picked from commit 5fc1dabfb3b2cbffdee3214d24d7769d6e440e45) Conflicts: src/msg/SimpleMessenger.cc commit 2f7979d1262e9d4899be76963a1620db46b334e8 Author: Samuel Just Date: Thu Jul 18 19:26:02 2013 -0700 ReplicatedPG: track temp collection contents, clear during on_change We also assert in on_flushed() that the temp collection is actually empty. Fixes: #5670 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 47516d9c4b7f023f3a16e166749fa7b1c7b3b24c) Conflicts: src/osd/ReplicatedPG.cc commit c7e2945a42541f966017180684dd969389eef3ac Author: Samuel Just Date: Thu Jul 18 19:25:14 2013 -0700 PG, ReplicatedPG: pass a transaction down to ReplicatedPG::on_change Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 9f56a7b8bfcb63cb4fbbc0c9b8ff01de9e518c57) commit 7ffc65fc4d7d842954cf791c016fd2711f644a9c Author: Samuel Just Date: Wed Jul 17 15:04:10 2013 -0700 PG: start flush on primary only after we process the master log Once we start serving reads, stray objects must have already been removed. Therefore, we have to flush all operations up to the transaction writing out the authoritative log. On replicas, we flush in Stray() if we will not eventually be activated and in ReplicaActive if we are in the acting set. This way a replica won't serve a replica read until the store is consistent. Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit b41f1ba48563d1d3fd17c2f62d10103b5d63f305) commit 850da0890da5df7e670df9268afe420d0c906c38 Author: Samuel Just Date: Wed Jul 17 12:51:19 2013 -0700 ReplicatedPG: replace clean_up_local with a debug check Stray objects should have been cleaned up in the merge_log transactions. Only on the primary have those operations necessarily been flushed at activate(). Fixes: 5084 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 278c7b59228f614addf830cb0afff4988c9bc8cb) commit 95b1b5da439f1b7e2fb1886aaeec2d61532183f0 Author: Samuel Just Date: Thu Jul 18 10:12:17 2013 -0700 FileStore: add global replay guard for split, collection_rename In the event of a split or collection rename, we need to ensure that we don't replay any operations on objects within those collections prior to that point. Thus, we mark a global replay guard on the collection after doing a syncfs and make sure to check that in _check_replay_guard() for all object operations. Fixes: #5154 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit f3f92fe21061e21c8b259df5ef283a61782a44db) Conflicts: src/os/FileStore.cc commit d92a43d8ff0123b234e47a94c2ce73fcaae7f625 Author: Samuel Just Date: Mon Jul 15 13:44:20 2013 -0700 OSD: add config option for peering_wq batch size Large peering_wq batch sizes may excessively delay peering messages resulting in unreasonably long peering. This may speed up peering. Backport: cuttlefish Related: #5084 Signed-off-by: Samuel Just Reviewed-by: Sage Weil (cherry picked from commit 39e5a2a406b77fa82e9a78c267b679d49927e3c3)