seems like a disk error, you can check your dmesg output, 2016-11-21 15:33 GMT+08:00 Peter Gervai <grin@xxxxxxx>: > Hello, > > This is Hammer(LTS), this may have been already fixed, maybe not, but > you have asked me to write to you, so I do. > > osd repeatedly failing (online for 2-3 days), with the same assert: > > 2016-11-21 02:20:26.115902 7f1756cb9700 -1 os/FileStore.cc: In > function 'virtual int FileStore::read(coll_t, const ghobject_t&, > uint64_t, size_t, ceph::bufferlist&, uint32_t, bool)' thread > 7f1756cb9700 time 2016-11-21 02:20:26.039891 > os/FileStore.cc: 2854: FAILED assert(allow_eio || > !m_filestore_fail_eio || got != -5) > > ceph version 0.94.9 (fe6d859066244b97b24f09d46552afc2071e6f90) > 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char > const*)+0x76) [0xc0f196] > 2: (FileStore::read(coll_t, ghobject_t const&, unsigned long, > unsigned long, ceph::buffer::list&, unsigned int, bool)+0xcc2) > [0x911012] > 3: (ReplicatedBackend::be_deep_scrub(hobject_t const&, unsigned int, > ScrubMap::object&, ThreadPool::TPHandle&)+0x31c) [0xa2268c] > 4: (PGBackend::be_scan_list(ScrubMap&, std::vector<hobject_t, > std::allocator<hobject_t> > const&, bool, unsigned int, > ThreadPool::TPHandle&)+0x2ca) [0x8d33fa] > 5: (PG::build_scrub_map_chunk(ScrubMap&, hobject_t, hobject_t, bool, > unsigned int, ThreadPool::TPHandle&)+0x1fa) [0x7dfdda] > 6: (PG::chunky_scrub(ThreadPool::TPHandle&)+0x3be) [0x7e835e] > 7: (PG::scrub(ThreadPool::TPHandle&)+0x1d7) [0x7e9a67] > 8: (OSD::ScrubWQ::_process(PG*, ThreadPool::TPHandle&)+0x19) [0x6b6ab9] > 9: (ThreadPool::worker(ThreadPool::WorkThread*)+0xa77) [0xbff747] > 10: (ThreadPool::WorkThread::entry()+0x10) [0xc00810] > 11: (()+0x80a4) [0x7f17794870a4] > 12: (clone()+0x6d) [0x7f17779df62d] > > I have the full event log on request. My problem is that I see no > other related log (with "perror" for example, as I have briefly seen > in the source, though I don't know where was it supposed to go), so I > don't know whether it's a disk error, a file format error or else. I > try to poke the osd with a 'ceph osd repair', and see what happens. > > Peter > -- > 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 -- Thank you! HuangJun -- 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
