Bug #4562: OSD fails to start with error assert(values.size() == 2) - Ceph - Ceph

Actions

Copy link

Bug #4562

closed

OSD fails to start with error assert(values.size() == 2)

Added by JuanJose Galvez about 11 years ago. Updated about 11 years ago.

Status:

Resolved

Priority:

Urgent

Assignee:

David Zafman

Category:

Target version:

% Done:

Source:

Development

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

The problem occurred after an upgrade to master (67c696dfac0a2343ec7530ca68e5ea9b2d3d31d0).

Error from log is:

2013-03-26 14:13:11.385596 7f7f1994e780 -1 osd/PG.cc: In function 'static int PG::read_info(ObjectStore*, coll_t, ceph::bufferlist&, pg_info_t&, std::map<unsigned int, pg_interval_t>&, hobject_t&, hobject_t&, interval_set<snapid_t>&, __u8&)' thread 7f7f1994e780 time 2013-03-26 14:13:11.384607
osd/PG.cc: 2785: FAILED assert(values.size() == 2)

The log is attached.

Files

ceph-osd.0.log (852 KB) ceph-osd.0.log

JuanJose Galvez, 03/26/2013 05:22 PM

Actions

Copy link

Updated by Sage Weil about 11 years ago

Priority changed from Normal to Urgent

Actions

Copy link

Updated by Ian Colle about 11 years ago

Assignee set to David Zafman

Actions

Copy link

Updated by David Zafman about 11 years ago

Status changed from New to In Progress

This was caused by the termination of an OSD while it was still running load_pgs() dealing with an upgrade from 0.56.4 to master.

It was found that PG:upgrade() includes a call to write_if_dirty() but with only dirty_info set. This would call write_info() upgrading the info_struct_v WITHOUT creating a new biginfo key.

The non-SnapMapper upgrade code path would normally call write_if_dirty() later when load_pgs() calls build_past_intervals_parallel().

Actions

Copy link