Bug #22631: mds: crashes because of old pool id in journal header - CephFS - Ceph

Actions

Copy link

Bug #22631

closed

mds: crashes because of old pool id in journal header

Added by dongdong tao over 6 years ago. Updated about 6 years ago.

Status:

Resolved

Priority:

Normal

Assignee:

dongdong tao

Category:

Target version:

% Done:

Source:

Tags:

Backport:

luminous,jewel

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Component(FS):

MDS, tools

Labels (FS):

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

we have use rados cppool command to copy the cephfs metadata pool
but,after copy done, mds would keep crashing when we restart it

home/jenkins-build/build/workspace/ceph-build/ARCH/x86_64/AVAILABLE_ARCH/x86_64/AVAILABLE_DIST/centos7/DIST/centos7/MACHINE_SIZE/huge/release/12.2.2/rpm/el7/BUILD/ceph-12.2.2/src/osdc/Journaler.cc: 87: FAILED assert(layout.pool_id pg_pool)
/home/jenkins-build/build/workspace/ceph-build/ARCH/x86_64/AVAILABLE_ARCH/x86_64/AVAILABLE_DIST/centos7/DIST/centos7/MACHINE_SIZE/huge/release/12.2.2/rpm/el7/BUILD/ceph-12.2.2/src/osdc/Journaler.cc: In function 'void Journaler::_set_layout(const file_layout_t*)' thread 7ff0a0595700 time 2018-01-09 02:26:47.129658
/home/jenkins-build/build/workspace/ceph-build/ARCH/x86_64/AVAILABLE_ARCH/x86_64/AVAILABLE_DIST/centos7/DIST/centos7/MACHINE_SIZE/huge/release/12.2.2/rpm/el7/BUILD/ceph-12.2.2/src/osdc/Journaler.cc: 87: FAILED assert(layout.pool_id pg_pool)
ceph version 12.2.2 (cf0baeeeeba3b47f9427c6c97e2144b094b7e5ba) luminous (stable)
1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0x110) [0x5556f751f850]
2: (()+0x544abf) [0x5556f748eabf]
3: (Journaler::_finish_read_head(int, ceph::buffer::list&)+0x2b5) [0x5556f7491e35]
4: (Context::complete(int)+0x9) [0x5556f71ff069]
5: (Finisher::finisher_thread_entry()+0x198) [0x5556f751e7c8]
6: (()+0x7e25) [0x7ff0ab7f3e25]
7: (clone()+0x6d) [0x7ff0aa8d634d]

Related issues 2 (0 open — 2 closed)

Actions

Copy link

Updated by dongdong tao over 6 years ago

through the code, we found it is because of the old pool id in the journal header.
my solution is
add "set pool_id" option to cephfs-journal-tool header mode, and it worked.

Actions

Copy link

Updated by dongdong tao over 6 years ago

https://github.com/ceph/ceph/pull/19860

Actions

Copy link

Updated by Jos Collin over 6 years ago

Status changed from New to Fix Under Review
Assignee set to dongdong tao

Actions

Copy link

Updated by Patrick Donnelly over 6 years ago

Subject changed from mds keep crashing because of old pool id in journal header to mds: crashes because of old pool id in journal header
Status changed from Fix Under Review to Pending Backport
Backport set to luminous,jewel
Component(FS) MDS, tools added