Project

General

Profile

Bug #46765

mds: segv in MDCache::wait_for_uncommitted_fragments

Added by Patrick Donnelly over 3 years ago. Updated over 3 years ago.

Status:
Resolved
Priority:
Urgent
Category:
-
Target version:
% Done:

0%

Source:
Q/A
Tags:
Backport:
octopus,nautilus
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
MDS
Labels (FS):
crash
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

2020-07-29T02:01:09.512+0000 7f058e203700  1 mds.0.27  waiting for uncommitted fragments
2020-07-29T02:01:09.512+0000 7f058e203700 -1 *** Caught signal (Segmentation fault) **
 in thread 7f058e203700 thread_name:MR_Finisher

 ceph version 16.0.0-3907-g30aa5304957 (30aa53049577e70968a728cf36a545a209e8d28b) pacific (dev)
 1: (()+0x128a0) [0x7f059bc028a0]
 2: (pthread_mutex_lock()+0) [0x7f059bbf9fa0]
 3: (C_GatherBase<MDSContext, C_MDSInternalNoop>::new_sub()+0x39) [0x55c2eac265f9]
 4: (MDCache::wait_for_uncommitted_fragments(C_GatherBase<MDSContext, C_MDSInternalNoop>*)+0x40) [0x55c2eab79c70]
 5: (MDSRank::rejoin_done()+0x2a7) [0x55c2eaab60d7]
 6: (MDSContext::complete(int)+0x52) [0x55c2ead7e582]
 7: (MDCache::open_snaprealms()+0xa53) [0x55c2eabddfb3]
 8: (MDCache::rejoin_gather_finish()+0x121) [0x55c2eabe4db1]
 9: (MDCache::rejoin_open_sessions_finish(std::map<client_t, std::pair<Session*, unsigned long>, std::less<client_t>, std::allocator<std::pair<client_t const, std::pair<Session*, unsigned long> > > >&)+0xea) [0x55c2eabf426a]
 10: (MDSContext::complete(int)+0x52) [0x55c2ead7e582]
 11: (MDSIOContextBase::complete(int)+0x233) [0x55c2ead7ea33]
 12: (MDSLogContextBase::complete(int)+0x40) [0x55c2ead7efa0]
 13: (Finisher::finisher_thread_entry()+0x195) [0x7f059c33fd55]
 14: (()+0x76db) [0x7f059bbf76db]
 15: (clone()+0x3f) [0x7f059addda3f]
 NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.

From: /ceph/teuthology-archive/pdonnell-2020-07-28_23:57:49-kcephfs-wip-pdonnell-testing-20200728.221932-distro-basic-smithi/5265360/remote/smithi186/log/ceph-mds.a.log.gz


Related issues

Related to CephFS - Bug #45699: mds may start to fragment dirfrag before rollback finishes Resolved
Copied to CephFS - Backport #46942: octopus: mds: segv in MDCache::wait_for_uncommitted_fragments Resolved
Copied to CephFS - Backport #46943: nautilus: mds: segv in MDCache::wait_for_uncommitted_fragments Resolved

History

#1 Updated by Patrick Donnelly over 3 years ago

  • Status changed from In Progress to Fix Under Review
  • Pull request ID set to 36358

#2 Updated by Patrick Donnelly over 3 years ago

  • Related to Bug #45699: mds may start to fragment dirfrag before rollback finishes added

#3 Updated by Patrick Donnelly over 3 years ago

  • Status changed from Fix Under Review to Pending Backport

#4 Updated by Nathan Cutler over 3 years ago

  • Copied to Backport #46942: octopus: mds: segv in MDCache::wait_for_uncommitted_fragments added

#5 Updated by Nathan Cutler over 3 years ago

  • Copied to Backport #46943: nautilus: mds: segv in MDCache::wait_for_uncommitted_fragments added

#6 Updated by Nathan Cutler over 3 years ago

  • Status changed from Pending Backport to Resolved

While running with --resolve-parent, the script "backport-create-issue" noticed that all backports of this issue are in status "Resolved" or "Rejected".

Also available in: Atom PDF