Bug #4927
closedOSD: pg upgrade does not clear snap_collections and split still iterates over snap_collections
0%
Updated by Samuel Just almost 11 years ago
(7812'4613,7812'5260] local-les=8406 n=421 ec=1 les/c 8406/8406 8217/8405/8405) [3,12] r=0 lpr=9302 lcod 0'0 mlcod 0'0 inactive] state<Start>: transitioning to Primary
6> 2013-05-07 19:30:02.175586 7f018ccb7700 2 filestore(/data/osd.3) _split_collection: 2.149_1a6 DNE 0.0.0.0:6801/17584 --> 149.165.228.133:6816/7350 -- pg_query(1.29a epoch 9303) v2 -- ?+0 0xc079c20 con 0xa0b0b00
-5> 2013-05-07 19:30:02.175629 7f0181693700 1 -
4> 2013-05-07 19:30:02.176225 7f017cd82700 2 - 0.0.0.0:6801/17584 >> 149.165.228.134:6804/6190 pipe(0x1e7ec80 sd=36 :0 s=1 pgs=0 cs=1 l=0).connect error 149.165.228.134:6804/6190, 111: Connection refused
3> 2013-05-07 19:30:02.176246 7f017cd82700 2 - 0.0.0.0:6801/17584 >> 149.165.228.134:6804/6190 pipe(0x1e7ec80 sd=36 :0 s=1 pgs=0 cs=1 l=0).fault 111: Connection refused
2> 2013-05-07 19:30:02.176313 7f017d085700 2 - 0.0.0.0:6801/17584 >> 149.165.228.133:6819/7915 pipe(0x1e7ef00 sd=32 :0 s=1 pgs=0 cs=1 l=0).connect error 149.165.228.133:6819/7915, 111: Connection refused
1> 2013-05-07 19:30:02.176327 7f017d085700 2 - 0.0.0.0:6801/17584 >> 149.165.228.133:6819/7915 pipe(0x1e7ef00 sd=32 :0 s=1 pgs=0 cs=1 l=0).fault 111: Connection refused
0> 2013-05-07 19:30:02.177552 7f018ccb7700 -1 os/FileStore.cc: In function 'int FileStore::_split_collection(coll_t, uint32_t, uint32_t, coll_t, const SequencerPosition&)' thread 7f018ccb7700 time 2013-05-07 19:30:02.175593
os/FileStore.cc: 4793: FAILED assert(replaying)
ceph version 0.61 (237f3f1e8d8c3b85666529860285dcdffdeda4c5)
1: /usr/bin/ceph-osd() [0x7eb334]
2: (FileStore::_do_transaction(ObjectStore::Transaction&, unsigned long, int)+0x4d3) [0x7fa8d3]
3: (FileStore::_do_transactions(std::list<ObjectStore::Transaction*, std::allocator<ObjectStore::Transaction*> >&, unsigned long, ThreadPool::TPHandle*)+0x69) [0x7fe3e9]
4: (FileStore::_do_op(FileStore::OpSequencer*, ThreadPool::TPHandle&)+0x17e) [0x7fe57e]
5: (ThreadPool::worker(ThreadPool::WorkThread*)+0x68a) [0x8e21fa]
6: (ThreadPool::WorkThread::entry()+0x10) [0x8e34a0]
7: (()+0x7e9a) [0x7f0196a94e9a]
8: (clone()+0x6d) [0x7f01952c6ccd]
NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.
Updated by Samuel Just almost 11 years ago
9303/8148) [] r=0 lpr=0 pi=8144-9302/4 lcod 0'0 mlcod 0'0 inactive] read_log done
0> 2013-05-07 18:58:40.324141 7f50385fd780 1 osd/OSD.cc: In function 'void OSD::load_pgs()' thread 7f50385fd780 time 2013-05-07 18:58:40.322945>second.empty())
osd/OSD.cc: 1779: FAILED assert(i
ceph version 0.61 (237f3f1e8d8c3b85666529860285dcdffdeda4c5)
1: (OSD::load_pgs()+0x2254) [0x6eb0f4]
2: (OSD::init()+0xc6e) [0x6ece9e]
3: (main()+0x2148) [0x621108]
4: (__libc_start_main()+0xed) [0x7f50361f376d]
5: /usr/bin/ceph-osd() [0x6235e9]
NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.
Updated by Mike Lowe almost 11 years ago
OSD's are able to start with wip_split_upgrade