Project

General

Profile

Actions

Bug #37786

closed

test fails in mon/crush_ops.sh

Added by Kefu Chai over 5 years ago. Updated over 4 years ago.

Status:
Can't reproduce
Priority:
High
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

2019-01-03T14:59:36.830 INFO:tasks.workunit.client.0.smithi144.stderr:+ ceph osd crush rm-device-class osd.0
...
2019-01-03T14:59:38.271 INFO:tasks.workunit.client.0.smithi144.stderr:osd.0 belongs to no class,
...
2019-01-03T14:59:59.242 INFO:tasks.workunit.client.0.smithi144.stderr:+ ceph osd crush set-device-class fish osd.0
...
2019-01-03T15:00:00.389 INFO:tasks.workunit.client.0.smithi144.stderr:Error EBUSY: osd.0 has already bound to class 'ssd', can not reset class to 'fish'; use 'ceph osd crush rm-device-class <id>' to remove old class first

see http://pulpito.ceph.com/kchai-2019-01-03_11:27:07-rados-wip-kefu2-testing-2019-01-03-1552-distro-basic-smithi/3417719/


Related issues 1 (0 open1 closed)

Has duplicate RADOS - Bug #49212: mon/crush_ops.sh fails: Error EBUSY: osd.1 has already bound to class 'ssd', can not reset class to 'hdd'Resolved

Actions
Actions #1

Updated by Kefu Chai over 5 years ago

2019-01-03 14:59:47.805 7f8c545b4700 10 mon.a@0(leader).paxosservice(osdmap 74..143) dispatch 0x55f4f7e39680 mon_command({"prefix": "osd crush rm-device-class", "ids": ["osd.0"]} v
0) v1 from client.114547 172.21.15.144:42578/3331020764 con 0x55f4f7888480
2019-01-03 14:59:47.805 7f8c545b4700 10 mon.a@0(leader).paxosservice(osdmap 74..143)  discarding forwarded message from previous election epoch 76 < 77

weird, the peon mon forwarded the message again, and the forwarded message was dropped by the leader.

Actions #2

Updated by Greg Farnum over 5 years ago

  • Priority changed from Normal to High

This does look weird to me as well; we should investigate. I can't offhand think of any way for a bug like that to impact only a subset of commands, but it might be that it only hits user commands or something...

Actions #3

Updated by Greg Farnum over 4 years ago

  • Status changed from New to Can't reproduce
Actions #4

Updated by Deepika Upadhyay about 3 years ago

  • Related to Bug #49212: mon/crush_ops.sh fails: Error EBUSY: osd.1 has already bound to class 'ssd', can not reset class to 'hdd' added
Actions #5

Updated by Deepika Upadhyay about 3 years ago

  • Related to deleted (Bug #49212: mon/crush_ops.sh fails: Error EBUSY: osd.1 has already bound to class 'ssd', can not reset class to 'hdd')
Actions #6

Updated by Deepika Upadhyay about 3 years ago

  • Has duplicate Bug #49212: mon/crush_ops.sh fails: Error EBUSY: osd.1 has already bound to class 'ssd', can not reset class to 'hdd' added
Actions

Also available in: Atom PDF