Bug #41514: in-flight manifest ops not properly cancelled on interval changing - RADOS - Ceph

Actions

Copy link

Bug #41514

closed

in-flight manifest ops not properly cancelled on interval changing

Added by xie xingguo over 4 years ago. Updated over 4 years ago.

Status:

Resolved

Priority:

Normal

Assignee:

Category:

Target version:

% Done:

Source:

Community (dev)

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

Ceph - v15.0.0

ceph-qa-suite:

rados

Component(RADOS):

Pull request ID:

29985

Crash signature (v1):

Crash signature (v2):

Description

which as a result makes PrimaryLogPG::on_flushed() unhappy:

/home/jenkins-build/build/workspace/ceph-dev-new-build/ARCH/x86_64/AVAILABLE_ARCH/x86_64/AVAILABLE_DIST/centos7/DIST/centos7/MACHINE_SIZE/huge/relea
se/15.0.0-4279-g09b93a5/rpm/el7/BUILD/ceph-15.0.0-4279-g09b93a5/src/osd/PrimaryLogPG.cc: 11829: FAILED ceph_assert(object_contexts.empty())

 ceph version 15.0.0-4279-g09b93a5 (09b93a5195fa25cb901a264c03478561aa89dfc5) octopus (dev)
 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0x14a) [0x5593658662ef]
 2: (()+0x4db4b7) [0x5593658664b7]
 3: (PrimaryLogPG::on_flushed()+0x1fe) [0x559365a5cd3e]
 4: (std::_Sp_counted_ptr_inplace<FlushState, std::allocator<FlushState>, (__gnu_cxx::_Lock_policy)2>::_M_dispose()+0x81) [0x559365a0aeb1]
 5: (()+0x6018a7) [0x55936598c8a7]
 6: (ContainerContext<std::shared_ptr<FlushState> >::~ContainerContext()+0x20) [0x5593659fdf40]
 7: (OSD::ShardedOpWQ::_process(unsigned int, ceph::heartbeat_handle_d*)+0x7bc) [0x559365963cbc]
 8: (ShardedThreadPool::shardedthreadpool_worker(unsigned int)+0x5b6) [0x559365f09836]
 9: (ShardedThreadPool::WorkThreadSharded::entry()+0x10) [0x559365f0b990]
 10: (()+0x7dd5) [0x7f3c6f132dd5]
 11: (clone()+0x6d) [0x7f3c6dff902d]