Bug 1324231

Summary:	[ceph-ansible] : purge cluster process should not check for OSD devices on MON and rgw nodes and should not give error 'Device not defined'
Product:	[Red Hat Storage] Red Hat Storage Console	Reporter:	Rachana Patel <racpatel>
Component:	ceph-ansible	Assignee:	Sébastien Han <shan>
Status:	CLOSED ERRATA	QA Contact:	Vasishta <vashastr>
Severity:	medium	Docs Contact:
Priority:	unspecified
Version:	2	CC:	adeza, aschoen, ceph-eng-bugs, gmeno, hnallurv, kdreyer, nthomas, racpatel, sankarshan, seb, shan, tchandra, uboppana
Target Milestone:	rc
Target Release:	2
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:	ceph-ansible-2.2.1-1.el7scon	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2017-06-19 13:14:42 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Rachana Patel 2016-04-05 22:04:12 UTC

Description of problem:
=======================
purging of cluster looks for OSd on mon and rgw nodes and gives error as below

TASK: [check for a device list] *********************************************** 
fatal: [magna042] => error while evaluating conditional: devices is not defined and osd_auto_discovery
fatal: [magna051] => error while evaluating conditional: devices is not defined and osd_auto_discovery
fatal: [magna063] => error while evaluating conditional: devices is not defined and osd_auto_discovery
skipping: [magna066]
skipping: [magna074]
fatal: [magna084] => error while evaluating conditional: devices is not defined and osd_auto_discovery
skipping: [magna067]


Version-Release number of selected component (if applicable):
=============================================================
ceph - 10.1.0-1.el7cp.x86_64
ceph-ansible-1.0.3-1.el7.noarch


How reproducible
================
always


Steps to Reproduce:
===================
1.create ceph cluster.
2. purge cluster using 'ansible-playbook purge-cluster.yml'


Actual results:
==============
TASK: [check for a device list] *********************************************** 
fatal: [magna042] => error while evaluating conditional: devices is not defined and osd_auto_discovery
fatal: [magna051] => error while evaluating conditional: devices is not defined and osd_auto_discovery
fatal: [magna063] => error while evaluating conditional: devices is not defined and osd_auto_discovery
skipping: [magna066]
skipping: [magna074]
fatal: [magna084] => error while evaluating conditional: devices is not defined and osd_auto_discovery
skipping: [magna067]

TASK: [get osd numbers] ******************************************************* 
ok: [magna074]
ok: [magna067]
ok: [magna066]



Expected results:
=================
no need to check for OSD device on mon/rgw node and it should not display error it as error -  'device not defined'  


Additional info:
================
[mons]
magna042
magna051
magna063

[osds]
magna066
magna067
magna074

[rgws]
magna084

Comment 2 Christina Meno 2016-04-25 21:48:01 UTC

I think what this error message is saying is that you must specify which device you used for the OSDs since guessing wrong could be disastrous.

Would you please try again? using something like 

--extra-vars '{"devices": ["/dev/YOUR_DEVICE_HERE"]}' 

when invoking the purge?

Comment 3 Rachana Patel 2016-08-07 17:05:31 UTC

Gregory, I understand we should give devices for OSD node but bug here is about MON nodes.
As i mentioned in Additional Info - magna042, magna051, magna063 are MON nodes

and error is 
'ASK: [check for a device list] *********************************************** 
fatal: [magna042] => error while evaluating conditional: devices is not defined and osd_auto_discovery'

So I think it should not complain for MON nodes

( Complain about one OSD node after that is acceptable)

Comment 4 Christina Meno 2016-09-21 21:41:24 UTC

please link upstream issue

Comment 7 seb 2016-09-22 14:22:55 UTC

ansible version?

Comment 10 Ken Dreyer (Red Hat) 2017-01-12 17:44:11 UTC

After further discussion with Andrew, we are not actually testing purging an RGW node in CI. Our CI system only tests purging 1-OSD 1-MON cluster, and we can't guarantee that purging anything else works properly today.

Comment 11 Andrew Schoen 2017-03-03 16:02:27 UTC

We now test purging a larger cluster upstream consisting of 3 MONs, 1 OSD, 1 MDS, 1 RGW and 1 Client node. I believe this ticket should be resolved now.

Comment 14 Tejas 2017-05-05 11:34:16 UTC

Verified in build:
ceph-ansible-2.2.4-1.el7scon.noarch

Comment 16 errata-xmlrpc 2017-06-19 13:14:42 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2017:1496