Bug 1528663

Summary:	After Openstack 10 triggers an "unknown" state on instances, when it recovers Cloudforms duplicates vms instead of recovering them
Product:	Red Hat CloudForms Management Engine	Reporter:	Felix Dewaleyne <fdewaley>
Component:	Providers	Assignee:	Sam Lucidi <slucidi>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Ido Ovadia <iovadia>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	5.8.0	CC:	agrare, akaris, bsorota, cpelland, fdewaley, gblomqui, gekis, igortiunov, jfrey, jhajyahy, jhardy, lavenel, lsmola, maufart, nchugh, niroy, obarenbo, ohadlevy, rspagnol, saali, simaishi, slucidi, smallamp
Target Milestone:	GA	Keywords:	TestOnly
Target Release:	5.10.0
Hardware:	All
OS:	All
Whiteboard:
Fixed In Version:	5.10.0.0	Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:
Clones:	1541042 1541072 (view as bug list)		Environment:
Last Closed:	2019-02-11 14:05:05 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	Openstack	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1541042, 1541072

Description Felix Dewaleyne 2017-12-22 17:18:28 UTC

Description of problem:
After Openstack 10 triggers an "unknown" state on instances, when it recoveres Cloudforms duplicates vms instead of recovering them

Version-Release number of selected component (if applicable):
openstack 10 (virtualised controllers)
Cloudforms 4.5.2

How reproducible:
every day at customer site

Steps to Reproduce:
1. get openstack to change instances to an unknwon state then back to normal
2. 
3.

Actual results:
the instances affected get archived and then get re-discovered

Expected results:
cloudforms recovers and does either un-archive the instances or never archives them in the first place

Additional info:
this happens daily at the customer's.

Comment 17 Adam Grare 2018-01-02 13:53:05 UTC

Hey Marek, it looks like some servers were returned multiple times.

Can you make a change to the openstack refresh_parser to unique the vms returned?
Adding this to the vcr tests would be great, also to confirm that graph refresh isn't affected since this should already be handled internally by graph refresh.

Comment 49 Jad Haj Yahya 2018-12-03 09:52:43 UTC

Hi Marek,

I need a clarification regarding step 2 above:

2. somehow filter some of VMs from Nova APIU output

Thanks,
Jad

Comment 50 Sam Lucidi 2018-12-04 21:41:34 UTC

Hi Jad,

I think all you should have to do to test whether the duplication has been solved is to add the provider, let it collect inventory, and then remove the provider so that the instances remain but are disconnected. Then readd the provider with all the same details and refresh to see whether the disconnected instances are reenabled or duplicated.

Sam

Comment 51 Jad Haj Yahya 2018-12-05 14:53:32 UTC

Verified on 5.10.0.27