Bug 1638303

Summary: The newly added compute node has domain part of FQDN missing in OSP services configuration
Product: Red Hat OpenStack Reporter: ojanas
Component: openstack-tripleo-heat-templatesAssignee: Alex Schultz <aschultz>
Status: CLOSED DUPLICATE QA Contact: Gurenko Alex <agurenko>
Severity: high Docs Contact:
Priority: high    
Version: 10.0 (Newton)CC: aschultz, dbecker, mburns, molasaga, morazi, sathlang, sputhenp
Target Milestone: ---Keywords: ZStream
Target Release: 10.0 (Newton)   
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: 2018-10-22 12:52:51 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:

Comment 2 Alex Schultz 2018-10-15 15:43:25 UTC
Tracked this down to https://review.openstack.org/#/c/563650/ which seems to be missing in the downstream.

Comment 6 Alex Schultz 2018-10-17 19:41:19 UTC
I tested the previous patch and that wasn't the fix. It appears that this issue is a duplicate of Bug 1600178.  Specifically because scaling up hits this code block to set the hostname:

https://github.com/openstack/puppet-tripleo/blob/newton-eol/manifests/profile/base/nova.pp#L93-L103

Moving over to the appropriate team for confirmation.

Comment 7 Sofer Athlan-Guyot 2018-10-22 12:52:51 UTC
Hi,

so this is a "duplicate" of https://bugzilla.redhat.com/show_bug.cgi?id=1600178, duplicate as we need the exact same patch to fix it.

The package puppet-tripleo-5.6.8-17.el7ost can be delivered as a hot fix and for scale out operations will have to be added the customer images before scale out.

To summarize the issue, we had to make sure for osp9/10 upgrade that we were keeping the same /etc/{nova/neutron}/{nova,neutron}.conf/host parameter value as we were switching from undefined parameter which was taking the default return by get_hostname python library and could be the short name to a defined value set to fqdn.

So we calculated the previous value and override the fqdn if it was the short name.

One unforseen side effect is that during OSP10 update (and also during scale out) we were also reverting it back to what it could have been in osp9.  The patch in 1600178 fixes that.

Don't hesitate to re-open it if the fixe in 1600178 applied to the image wouldn't solve the issue here.

Regards,

*** This bug has been marked as a duplicate of bug 1600178 ***