Bug #18303: osd_backfill.py: ENXIO on flush_pg_stats - Ceph - Ceph

Actions

Copy link

Bug #18303

closed

osd_backfill.py: ENXIO on flush_pg_stats

Added by Sage Weil over 7 years ago. Updated over 7 years ago.

Status:

Resolved

Priority:

Urgent

Assignee:

Category:

Target version:

% Done:

Source:

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

2016-12-20T01:23:07.870 INFO:tasks.ceph.osd.1.smithi100.stdout:starting osd.1 at - osd_data /var/lib/ceph/osd/ceph-1 /var/lib/ceph/osd/ceph-1/journal
2016-12-20T01:23:07.872 INFO:tasks.ceph.osd.1.smithi100.stderr:2016-12-20 01:23:07.872138 7f8c8b2caa40 -1 WARNING: the following dangerous and experimental features are enabled: *
,,,
2016-12-20T01:23:08.858 INFO:tasks.ceph.osd.1.smithi100.stderr:2016-12-20 01:23:08.857947 7f8c8b2caa40 -1 osd.1 16 log_to_monitors {default=true}
...
2016-12-20T01:23:12.903 INFO:teuthology.orchestra.run.smithi100:Running: 'sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 0 ceph --cluster ceph --admin-daemon /var/run/ceph/ceph-osd.1.asok dump_ops_in_flight'
...
2016-12-20T01:23:13.771 INFO:tasks.osd_backfill:err is 0
...
2016-12-20T01:23:13.771 INFO:teuthology.orchestra.run.smithi100:Running: 'sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 120 ceph --cluster ceph tell osd.1 flush_pg_stats'
2016-12-20T01:23:13.880 INFO:teuthology.orchestra.run.smithi100.stderr:2016-12-20 01:23:13.880074 7f215341e700 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-12-20T01:23:13.891 INFO:teuthology.orchestra.run.smithi100.stderr:2016-12-20 01:23:13.890621 7f215341e700 -1 WARNING: the following dangerous and experimental features are enabled: *
2016-12-20T01:23:16.056 INFO:teuthology.orchestra.run.smithi100.stderr:Error ENXIO: problem getting command descriptions from osd.1

The task waits for the io to succeed, but does not wait for osd.1 and osd.2 to be up. Probably only osd.2 is up and peered but the flush on osd.1 fails?