Bug 1389271

Summary:	Deployment fails if you skip introspection
Product:	Red Hat OpenStack	Reporter:	Udi Kalifon <ukalifon>
Component:	rhosp-director	Assignee:	Angus Thomas <athomas>
Status:	CLOSED WORKSFORME	QA Contact:	Omri Hochman <ohochman>
Severity:	urgent	Docs Contact:
Priority:	unspecified
Version:	10.0 (Newton)	CC:	dbecker, dsneddon, dtantsur, mburns, morazi, rhel-osp-director-maint, sasha, ukalifon
Target Milestone:	---
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2016-10-31 09:19:50 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:

Description Udi Kalifon 2016-10-27 10:25:15 UTC

Description of problem:
I am working on a bare metal setup, and when I register my nodes I already have all the correct RAM amount and disk space for every machine. In theory, this should enable me to skip the introspection step - but if I try that the deployment fails with "no valid host found".

What's also strange is that the deployment fails not right at the beginning... For about 10-15 minutes it seems like the deployment is working. You'd expect such an error to occur at a much earlier stage... In ironic you see the servers stuck in BUILD state (only 1 server is in ERROR state). I also checked with 'nova hypervisor-stats' before and after introspection and I see more or less the same resources available so it should have worked also when I skipped introspection.

What does introspection do that deployment fails without it? By the way, in 10 the registration of the nodes already makes the nodes 'available' so there's no risk that you didn't make them available.

Version-Release number of selected component (if applicable):
python-tripleoclient-5.3.0-2.el7ost.noarch

How reproducible:
100%

Steps to Reproduce:
1. Register nodes and make sure that instackenv.json contains the technical specs (RAM, CPU, disk) for each node
2. Use 'nova hypervisor-stats' to see more or less that the total amount of resources is what's expected
3. Run 'openstack baremetal configure boot' but don't run introspection.
4. Make sure all nodes are available
5. Deploy
6. If you fail, delete the stack and run introspection - then deploy with the same command again

Actual results:
You can't deploy if you skip introspection

Expected results:
Introspection should be an optional step

Additional info:

Comment 1 Dmitry Tantsur 2016-10-31 08:00:14 UTC

Hi!

Please provide some information about your case (instackenv.json, node details, ironic port list, flavor details, etc). It's impossible to say something specific without knowing what steps you took. Introspection is not mandatory, but you have to take care about filling in properties and creating ports yourself.

Comment 2 Udi Kalifon 2016-10-31 09:19:50 UTC

This case is working now. If I hit it again I'll be able to investigate more. Thanks for the information.