Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1669345

Summary: IPMI driver not working with IPv6 endpoints - affects the ability to deploy open stack in production for a project
Product: Red Hat OpenStack Reporter: Cristian Muresanu <cmuresan>
Component: openstack-ironic-inspectorAssignee: RHOS Maint <rhos-maint>
Status: CLOSED CURRENTRELEASE QA Contact: mlammon
Severity: high Docs Contact:
Priority: high    
Version: 13.0 (Queens)CC: acontant, bfournie, cmuresan, cpaquin, dmalea, jthomas, mburns, racedoro, rhos-maint, slinaber, sputhenp
Target Milestone: z6Keywords: TestOnly, Triaged, ZStream
Target Release: 13.0 (Queens)   
Hardware: Unspecified   
OS: Linux   
Whiteboard:
Fixed In Version: openstack-ironic-inspector-7.2.1-6.el7ost Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
: 1687986 (view as bug list) Environment:
Last Closed: 2019-05-24 20:59:47 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:    
Bug Blocks: 1687986    

Description Cristian Muresanu 2019-01-25 01:06:47 UTC 
Description of problem:
We need to run socat in order to translate IPv4 IPMI requests to IPv6 when communicating with IPMI endpoints that are running IPv6 only.  We do not want to use socat for this functionality as it is very hard to maintain.

Apparently there is a bug in the IPMI driver for Ironic

Version-Release number of selected component (if applicable):

Workaround by Joe Antkowiak:
As a workaround I set up a background forked socat process for each node, that forwarded a local UDP port via IPv4 to a remote IPv6 address on its IPv6 address.  This worked, but CU would like to not have to use this workaround in production and wants to get ironic IPv6 IPMI to work

How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:
It affects our ability to deploy open stack in production for a project.

Expected results:


Additional info:
Comment 2 Bob Fournier 2019-02-13 11:03:23 UTC 
It looks like this issue is being tracked upstream - https://storyboard.openstack.org/#!/story/2004648, with a patch proposed - https://review.openstack.org/#/c/626552/.

When the patch lands we'll get it back ported to Queens and downstream to OSP-13.
Comment 4 Bob Fournier 2019-03-08 14:59:08 UTC 
We also need this one - https://review.openstack.org/#/c/642039
Comment 11 mlammon 2019-04-16 14:26:33 UTC 
Christian, have you had a chance to test the hotfix?
Comment 13 Cristian Muresanu 2019-04-29 02:49:22 UTC 
From the CU side:
"We are testing the fixes now and they look good but I am waiting for our redhat account team to also confirm"
Comment 16 Cristian Muresanu 2019-05-01 15:27:20 UTC 
From the CU side/ticket:
"Sorry I missed this, but we've have had no issues with the hotfix since implementation on the newkirk stack."
Comment 17 mlammon 2019-05-06 18:11:53 UTC 
Based on comments -> https://bugzilla.redhat.com/show_bug.cgi?id=1669345#c16, I will mark this verified.
Comment 18 Bob Fournier 2019-05-24 20:59:47 UTC 
Closing as this is TestOnly.