Project

General

Profile

Bug #24324

[Hammer]pg trim got segmentation fault

Added by 相洋 于 7 months ago. Updated 7 months ago.

Status:
Won't Fix
Priority:
Normal
Assignee:
-
Category:
OSD
Target version:
Start date:
05/28/2018
Due date:
% Done:

0%

Source:
Support
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:

Description

Last Friday, we got a osd down from our monitoring system, we got the osd log as below.

Does anyone encouter this problem? Any clue?

2018-05-25 21:12:04.566736 7fee7b9ec700 -1 ** Caught signal (Segmentation fault) *in thread 7fee7b9ec700
ceph version 0.94.5 (b446846bba51e5e5753749b33d4cfe8134544c32)
1: /usr/bin/ceph-osd() [0xa4c7d5]

2: /lib64/libpthread.so.0 [0x3cda00f7e0]
3. /usr/lib64/libjemalloc.so.1() [0x0x3c1c0a167]
4. /usr/lib64/libjemalloc.so.1() [0x0x3c1c0e8d9]
5. /usr/lib64/libjemalloc.so.1() [0x0x3c1c27dec]
6. (free()+0x3c9) [0x3ce1c0a009]
7. (PGlog::IndexedLog::trim(PGLog::LogEntryHandle*, eversion_t, std::set<eversion_t, std::less<eversion_t>, std::allocator<eversion_t>> ) +0x30c)[0x75535c]
8. (PGLog:trim(PGLog::LogEntryHandle
, eversion_t, pg_info_t&), +0x1ce) [0x755a9e]
9. (PG::apend_log(std::vector<pg_log_entry_t, std::allocator<pg_log_entry_t>> consit&, eversion_t, eversion_t, ObjectStore::Transaction&, bool)+0x2da)[0x7e436a] 10. (ReplicatedPG::log_operation(std::vector<pg_log_entry_t, std::allocator<pg_log_entry_t>> const &,boost::optional<pg_hit_set_history_t> &, eversion_t const &,eversion_t const &,bool,ObjectStore::Transaction ) + 0xd1)[0x8a5851]
11: (void ReplicatedBackend::sub_op_modify_impl<MOSDRepOp, 112>(std::tr1::shared)ptr<OpRequest>)+0x8bb) [0x9df5cb]
12: (ReplicatedBackend::sub_op_modify(std::shared_ptr<OpRequest>)+0x953) [0x55e696dff453]
13: (ReplicatedBackend::handle_message(std::shared_ptr<OpRequest>)+0x3b3) [0x55e696e00003]
14: (PrimaryLogPG::do_request(std::shared_ptr<OpRequest>&, ThreadPool::TPHandle&)+0xed) [0x55e696ccb39d]
15: (OSD::dequeue_op(boost::intrusive_ptr<PG>, std::shared_ptr<OpRequest>, ThreadPool::TPHandle&)+0x3f5) [0x55e696b7e255]
16: (OSD::ShardedOpWQ::_process(unsigned int, ceph::heartbeat_handle_d
)+0x7dd) [0x55e696ba4c0d]
17: (ShardedThreadPool::shardedthreadpool_worker(unsigned int)+0x947) [0x55e6971d0af7]
18: (ShardedThreadPool::WorkThreadSharded::entry()+0x10) [0x55e6971d2c50]
19: +/lib64/libpthread.so.0 [0x7feec181d184]
20: (clone()+0x6d) [0x7feebff7137d]
NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.

History

#1 Updated by Sage Weil 7 months ago

  • Status changed from New to Won't Fix

This crash doesn't look familiar.

Since hammer is end-of-life we don't have time to look into the crash. I'd recommend upgrading to jewel or luminous.

Also available in: Atom PDF