I just upgraded to Ceph 0.24, and when I started up the cluster I got this crash. 2011-01-06 21:23:48.744864 b6855b70 mds0.cache creating system inode with ino:601 2011-01-06 21:23:48.745488 b6855b70 log [ERR] : unmatched fragstat size on single dirfrag 100, inode has f(v0 m2011-01-06 21:23:48.745063 1=0+1), dirfrag has f(v0 m2011-01-06 21:23:48.745063 3=1+2) 2011-01-06 21:23:49.293433 b6855b70 bad get [inode 600 [...2,head] ~mds0/stray/ auth v26003 f(v8833 m2011-01-04 21:50:03.135177 8846=8435+411) n(v11792 rc2011-01-04 21:50:03.135177 b25052311635 a-18 8846=8435+411) (inest lock dirty) (ifile lock dirty) (iversion lock) | dirtyscattered dirfrag stray dirty 0x11102a68] by 19 stray was 5 (-1005,-1005,-1,19,1001) mds/CInode.h: In function 'virtual void CInode::bad_get(int)': mds/CInode.h:1088: FAILED assert(ref_set.count(by) == 0) ceph version 0.24 (commit:180a4176035521940390f4ce24ee3eb7aa290632) 1: (CInode::bad_put(int)+0) [0x827b090] 2: (MDSCacheObject::get(int)+0x153) [0x813e463] 3: (MDCache::populate_mydir()+0x8a) [0x81a7e5a] 4: (MDCache::_create_system_file_finish(Mutation*, CDentry*, Context*)+0x181) [0x819f501] 5: (C_MDC_CreateSystemFile::finish(int)+0x29) [0x81d6c29] 6: (finish_contexts(std::list<Context*, std::allocator<Context*> >&, int)+0x6b) [0x81d663b] 7: (Journaler::_finish_flush(int, long long, utime_t, bool)+0x983) [0x82f2f53] 8: (Journaler::C_Flush::finish(int)+0x3f) [0x82fb24f] 9: (Objecter::handle_osd_op_reply(MOSDOpReply*)+0x801) [0x82d8e31] 10: (MDS::_dispatch(Message*)+0x2ae5) [0x80eaa15] 11: (MDS::ms_dispatch(Message*)+0x62) [0x80eb142] 12: (SimpleMessenger::dispatch_entry()+0x899) [0x80b8649] 13: (SimpleMessenger::DispatchThread::entry()+0x22) [0x80b30f2] 14: (Thread::_entry_func(void*)+0x11) [0x80c9101] 15: (()+0x5cc9) [0x6ffcc9] 16: (clone()+0x5e) [0x7e669e] NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this. (The objdump it mentions is attached.) Anybody seen this before? --Ravi -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
