Actions
Bug #17857
closedosd/osd_types.h: 4287: FAILED assert(rwstate.empty())
% Done:
0%
Source:
Development
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
ceph-helpers.sh:1103: wait_for_clean: returned 1
Updated by Kefu Chai over 7 years ago
ctest -R test-erasure-eio.sh
failed in TEST_rados_get_bad_size_shard_0
2016-11-11 15:23:07.863392 7f8f4bdef700 -1 /var/ceph/ceph/src/osd/osd_types.h: In function 'ObjectContext::~ObjectContext()' thread 7f8f4bdef700 time 2016-11-11 15:23:07.833415 /var/ceph/ceph/src/osd/osd_types.h: 4287: FAILED assert(rwstate.empty()) ceph version Development (no_version) 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0x95) [0x562ee684c720] 2: (ObjectContext::~ObjectContext()+0x41) [0x562ee609923b] 3: (SharedLRU<hobject_t, ObjectContext, hobject_t::ComparatorWithDefault, std::hash<hobject_t> >::Cleanup::operator()(ObjectContext*)+0x40) [0x562ee611431a] 4: (std::_Sp_counted_deleter<ObjectContext*, SharedLRU<hobject_t, ObjectContext, hobject_t::ComparatorWithDefault, std::hash<hobject_t> >::Cleanup, std::allocator<void>, (__gnu_c\xx::_Lock_policy)2>::_M_dispose()+0x33) [0x562ee6120315] 5: (std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release()+0x42) [0x562ee5d89930] 6: (std::__shared_count<(__gnu_cxx::_Lock_policy)2>::~__shared_count()+0x27) [0x562ee5d88e87] 7: (std::__shared_ptr<ObjectContext, (__gnu_cxx::_Lock_policy)2>::~__shared_ptr()+0x1c) [0x562ee5eed062] 8: (std::shared_ptr<ObjectContext>::~shared_ptr()+0x18) [0x562ee5eed0a4] 9: (SharedLRU<hobject_t, ObjectContext, hobject_t::ComparatorWithDefault, std::hash<hobject_t> >::clear()+0xd3) [0x562ee60bbc19] 10: (ReplicatedPG::on_change(ObjectStore::Transaction*)+0x96d) [0x562ee606c907] 11: (PG::start_peering_interval(std::shared_ptr<OSDMap const>, std::vector<int, std::allocator<int> > const&, int, std::vector<int, std::allocator<int> > const&, int, ObjectStore\ ::Transaction*)+0xea4) [0x562ee5f2bf6e] 12: (PG::RecoveryState::Reset::react(PG::AdvMap const&)+0x514) [0x562ee5f3288e]
Updated by Kefu Chai over 7 years ago
we failed to release the read lock on object being recovered
the read lock is acquired by obc->get_recovery_read()
, in ReplicatedPG::prep_object_replica_pushes()
,
Updated by Kefu Chai over 7 years ago
- Status changed from New to In Progress
- Assignee set to Kefu Chai
Updated by Kefu Chai over 7 years ago
- Status changed from In Progress to Fix Under Review
Updated by Loïc Dachary over 7 years ago
https://github.com/ceph/ceph/pull/11979/commits/8854cca4164f9184cc549ba0b90b44515933de8c disables test-erasure-eio.sh until this is fixed because it fails most of the time. To verify the fix is good, it is recommended to run the test in a loop on a machine that has a high load. For instance running a ./run-make-check.sh at the same time is enough to recreate conditions similar to jenkins on a machine such as rex001.
Updated by Kefu Chai over 7 years ago
- Subject changed from ceph-helpers.sh:1103: wait_for_clean: returned 1 to osd/osd_types.h: 4287: FAILED assert(rwstate.empty())
- Description updated (diff)
Updated by Kefu Chai over 7 years ago
- Has duplicate Bug #17860: osd/osd_types.h: 4287: FAILED assert(rwstate.empty()) added
Updated by Kefu Chai over 7 years ago
- Related to Bug #17830: osd-scrub-repair.sh is failing (intermittently?) on Jenkins added
Updated by Kefu Chai over 7 years ago
- Status changed from Fix Under Review to Resolved
Actions