Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 965787

Summary:	openstack-packstack: Installation failed on iptables Command Error 'Resource temporarily unavailable'.
Product:	Red Hat OpenStack	Reporter:	Jaroslav Henner <jhenner>
Component:	openstack-packstack	Assignee:	Martin Magr <mmagr>
Status:	CLOSED ERRATA	QA Contact:	Nir Magnezi <nmagnezi>
Severity:	high	Docs Contact:
Priority:	medium
Version:	3.0	CC:	ajeain, aortega, derekh, jhenner, mmagr, ohochman, sclewis, sgordon, ykaul
Target Milestone:	snapshot2	Keywords:	TestBlocker, Triaged
Target Release:	3.0
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:	openstack-packstack-2013.1.1-0.8.dev601	Doc Type:	Bug Fix
Doc Text:	Previously a race condition existed when deploying compute nodes whereby the order in which libvirt and PackStack itself modified the firewall rules was unpredictable. As a result the iptables firewall rules generated by the deployment would sometimes be incorrect, resulting in connectivity issues. PackStack has been updated to prevent this race condition from occurring and ensure the correct iptables firewall rules are deployed.	Story Points:	---
Clone Of:	908695	Environment:
Last Closed:	2013-06-11 18:51:23 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:
Bug Depends On:	908695
Bug Blocks:

Comment 1 Jaroslav Henner 2013-05-21 17:47:42 UTC

This problem occurs rarely, maybe 5% of deployments. Probably only when one is deploying a compute node on separate servers.

Comment 5 Martin Magr 2013-06-05 11:58:49 UTC

The problem was in case you had separate compute nodes in your setup. The race condition for firewall wasn't appended to manifestfor them.

So I suggest running setup with 1 controller node and 1+ compute nodes. You can also check in /var/tmp/packstack/<YYMMDD-hhmmss>-<random>/manifests/<compute_host_ip>_nova.pp.log that firewall rule '001 nova compute incoming' is applied before class 'nova::compute::libvirt' (or more precisely before service 'libvirt').

Comment 6 Nir Magnezi 2013-06-05 13:31:17 UTC

(In reply to Martin Magr from comment #5)
> The problem was in case you had separate compute nodes in your setup. The
> race condition for firewall wasn't appended to manifestfor them.
> 
> So I suggest running setup with 1 controller node and 1+ compute nodes. You
> can also check in
> /var/tmp/packstack/<YYMMDD-hhmmss>-<random>/manifests/<compute_host_ip>_nova.
> pp.log that firewall rule '001 nova compute incoming' is applied before
> class 'nova::compute::libvirt' (or more precisely before service 'libvirt').

Thank you!

Comment 7 Nir Magnezi 2013-06-05 13:34:23 UTC

Verified NVR: openstack-packstack-2013.1.1-0.8.dev601.el6ost.noarch

Installed Openstack via packstack with 1 controller and 2 compute nodes.
No errors were indicated in stdout nor in packstack logs.

Comment 8 errata-xmlrpc 2013-06-11 18:51:23 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.

http://rhn.redhat.com/errata/RHBA-2013-0938.html