Project

General

Profile

Bug #19771

HEALTH_WARN pool rbd pg_num 244 > pgp_num 224 during upgrade

Added by Sage Weil 28 days ago. Updated 19 days ago.

Status:
Pending Backport
Priority:
Immediate
Assignee:
Category:
-
Target version:
-
Start date:
04/25/2017
Due date:
% Done:

0%

Source:
Tags:
Backport:
kraken
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Release:
Needs Doc:
No

Description

2017-04-25T16:42:57.790 INFO:teuthology.orchestra.run.smithi078:Running: 'adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage ceph --cluster ceph health'
2017-04-25T16:42:57.929 INFO:teuthology.misc.health.smithi078.stdout:HEALTH_WARN pool rbd pg_num 244 > pgp_num 224
2017-04-25T16:42:57.929 DEBUG:teuthology.misc:Ceph health: HEALTH_WARN pool rbd pg_num 244 > pgp_num 224
2017-04-25T16:43:04.931 INFO:teuthology.orchestra.run.smithi078:Running: 'adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage ceph --cluster ceph health'
2017-04-25T16:43:05.093 INFO:teuthology.misc.health.smithi078.stdout:HEALTH_WARN pool rbd pg_num 244 > pgp_num 224

/a/sage-2017-04-25_15:37:28-upgrade:jewel-x-master---basic-smithi/1068352
description: upgrade:jewel-x/stress-split-erasure-code/{0-cluster/{openstack.yaml
  start.yaml} 1-jewel-install/jewel.yaml 2-partial-upgrade/firsthalf.yaml 3-thrash/default.yaml
  4-workload/ec-rados-default.yaml 5-finish-upgrade.yaml 6-luminous.yaml 7-final-workload/ec-rados-plugin=jerasure-k=3-m=1.yaml
  distros/ubuntu_14.04.yaml}

Related issues

Copied to Backport #20024: kraken: HEALTH_WARN pool rbd pg_num 244 > pgp_num 224 during upgrade New

History

#1 Updated by Sage Weil 22 days ago

/a/sage-2017-05-01_11:43:01-upgrade:kraken-x-master---basic-smithi/1087965

#2 Updated by Kefu Chai 21 days ago

2017-05-01T13:00:05.534 INFO:tasks.thrashosds.thrasher:fixing pg num pool rbd
...
{"pgid":"0.147","version":"0'0","reported_seq":"0","reported_epoch":"0","state":"creating","last_fresh":"2017-05-01 12:59:46.655825" 
...
{"pgid":"0.146","version":"0'0","reported_seq":"0","reported_epoch":"0","state":"creating","last_fresh":"2017-05-01 12:57:47.667230",

some pgs were still being created.

    def set_pool_pgpnum(self, pool_name):
        """ 
        Set pgpnum property of pool_name pool.
        """ 
        with self.lock:
            assert isinstance(pool_name, basestring)
            assert pool_name in self.pools
            if self.get_num_creating() > 0:
                return False
            self.set_pool_property(pool_name, 'pgp_num', self.pools[pool_name])
            return True


so we didn't fix the pg_num of pool rbd when finishing thrashing osd.

#3 Updated by Kefu Chai 21 days ago

  • Status changed from New to Need Review
  • Assignee set to Kefu Chai

#4 Updated by Sage Weil 19 days ago

  • Status changed from Need Review to Pending Backport
  • Backport set to kraken

#5 Updated by Nathan Cutler 1 day ago

  • Copied to Backport #20024: kraken: HEALTH_WARN pool rbd pg_num 244 > pgp_num 224 during upgrade added

Also available in: Atom PDF