Project

General

Profile

Bug #9173

Crash in Server::_session_logged

Added by John Spray over 9 years ago. Updated over 9 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
-
Target version:
-
% Done:

0%

Source:
other
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
Labels (FS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Hit by mds_client_recovery task

2014-08-20 14:13:00,479.479 INFO:tasks.ceph.mds.a.mira097.stderr:ceph-mds: /usr/include/boost/smart_ptr/intrusive_ptr.hpp:166: T* boost::intrusive_ptr<T>::operator->() const [with T = Connection]: Assertion `px != 0' failed.
2014-08-20 14:13:00,480.480 INFO:tasks.ceph.mds.a.mira097.stderr:*** Caught signal (Aborted) **
2014-08-20 14:13:00,480.480 INFO:tasks.ceph.mds.a.mira097.stderr: in thread 7f32fc112700
2014-08-20 14:13:00,485.485 INFO:tasks.ceph.mds.a.mira097.stderr: ceph version 0.84-365-g790de97 (790de974a88c1cd26ca5b96d85b6ed8a5a5bfa8b)
2014-08-20 14:13:00,486.486 INFO:tasks.ceph.mds.a.mira097.stderr: 1: ceph-mds() [0x810cfa]
2014-08-20 14:13:00,486.486 INFO:tasks.ceph.mds.a.mira097.stderr: 2: (()+0xfcb0) [0x7f3300a01cb0]
2014-08-20 14:13:00,486.486 INFO:tasks.ceph.mds.a.mira097.stderr: 3: (gsignal()+0x35) [0x7f32ff4f0425]
2014-08-20 14:13:00,486.486 INFO:tasks.ceph.mds.a.mira097.stderr: 4: (abort()+0x17b) [0x7f32ff4f3b8b]
2014-08-20 14:13:00,486.486 INFO:tasks.ceph.mds.a.mira097.stderr: 5: (()+0x2f0ee) [0x7f32ff4e90ee]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 6: (()+0x2f192) [0x7f32ff4e9192]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 7: ceph-mds() [0x57215d]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 8: (Server::_session_logged(Session*, unsigned long, bool, unsigned long, interval_set<inodeno_t>&, unsigned long)+0x151b) [0x5adceb]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 9: (C_MDS_session_finish::finish(int)+0x38) [0x5ff478]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 10: (Context::complete(int)+0x9) [0x595859]
2014-08-20 14:13:00,487.487 INFO:tasks.ceph.mds.a.mira097.stderr: 11: (finish_contexts(CephContext*, std::list<Context*, std::allocator<Context*> >&, int)+0x95) [0x597065]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 12: (Journaler::_finish_flush(int, unsigned long, utime_t)+0x1b2) [0x7a3b82]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 13: (Context::complete(int)+0x9) [0x595859]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 14: (Objecter::handle_osd_op_reply(MOSDOpReply*)+0x1073) [0x7be213]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 15: (MDS::handle_core_message(Message*)+0x5b8) [0x592948]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 16: (MDS::_dispatch(Message*)+0x2f) [0x59308f]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 17: (MDS::ms_dispatch(Message*)+0x1e4) [0x594af4]
2014-08-20 14:13:00,488.488 INFO:tasks.ceph.mds.a.mira097.stderr: 18: (Messenger::ms_deliver_dispatch(Message*)+0x77) [0x9c3f87]
2014-08-20 14:13:00,489.489 INFO:tasks.ceph.mds.a.mira097.stderr: 19: (DispatchQueue::entry()+0x4b2) [0x9c0d22]
2014-08-20 14:13:00,489.489 INFO:tasks.ceph.mds.a.mira097.stderr: 20: (DispatchQueue::DispatchThread::entry()+0xd) [0x8d5bcd]
2014-08-20 14:13:00,489.489 INFO:tasks.ceph.mds.a.mira097.stderr: 21: (()+0x7e9a) [0x7f33009f9e9a]
2014-08-20 14:13:00,489.489 INFO:tasks.ceph.mds.a.mira097.stderr: 22: (clone()+0x6d) [0x7f32ff5ae3fd]
2014-08-20 14:13:00,490.490 INFO:tasks.ceph.mds.a.mira097.stderr:2014-08-20 06:12:58.495556 7f32fc112700 -1 *** Caught signal (Aborted) **
2014-08-20 14:13:00,490.490 INFO:tasks.ceph.mds.a.mira097.stderr: in thread 7f32fc112700

ceph-mds.a.log View (51.8 KB) John Spray, 08/20/2014 06:30 AM

9173.mds.log View (1.46 MB) John Spray, 08/20/2014 10:47 AM


Related issues

Duplicated by CephFS - Bug #9297: mds: null deref in _session_logged Duplicate 08/30/2014

Associated revisions

Revision 493577bf (diff)
Added by John Spray over 9 years ago

mds: fix crash killing sessions without conn

Bug was introduced in 54eca56695 (mds: use Connection::mark_down()).
Used to call SimpleMessenger::mark_down, which had a check for
connection==NULL, when calling conn->mark_down() we must
now do our own nullness check.

Fixes: #9173

Signed-off-by: John Spray <>

History

#1 Updated by John Spray over 9 years ago

Better log.

#2 Updated by John Spray over 9 years ago

  • Status changed from New to Fix Under Review

#3 Updated by John Spray over 9 years ago

  • Status changed from Fix Under Review to Resolved

Also available in: Atom PDF