Missing cluster nodes in vRA's VAMI after 7.3 upgrade

[defmania]

Hello,

I have a problem after upgrading an enterprise deployment from 7.2 to 7.3. All the nodes are not showing up on the VAMI's vRA Settings - Cluster page.

The upgrade was successful and everything is working in vRA, provisioning works, nodes are showing up in the web UI on the DEM Status page, but they are missing from the VAMI.

Any thoughts?

Thanks!

[darrengoff3]

Have you tried re-installing the IaaS Management agents on the IaaS nodes? I had nodes not connecting after an upgrade and needed to re-install the IaaS Management Agents. Did you have any problems during your upgrade? Is the screenshot from your primary vRA Appliance?

[defmania]

Yes. I've reinstalled the management agent on one of the DEMs and still no luck in making it show up on the VAMI. I've checked all the management agents versions and they are all on 7.3.

There were no problems during the upgrade, everything was ok. The screenshot is from the primary node, but it's looking exactly the same on the secondary node.

[darrengoff3]

Have you performed a clean startup in the correct sequence since the upgrade? Are there any errors in the /var/log/vcac/catalina.out log on the vRA Appliance?

[darrengoff3]

Are the DB nodes present on the database tab?

[defmania]

Yes, I've just performed another clean "shutdown" - "power on" for all the nodes and they are still missing from the VAMI.

DB nodes are looking ok. There are a few errors showing up in the catalina.out log, but they appear to be related to other things:

"com.vmware.vcac.iaas.service.impl.MachineResourceActionServiceImpl.loadIcon:273 - Exception during loading resource action icon with url: icon_vco.png"

Thanks!

[darrengoff3]

Have the services started on the vRA Appliances?

[gradinka]

one thing worth check is the configuration of management agents, in their xml config files on the IaaS machines;

check what are they pointing at - primary VA, replica VA, load balancer or that is it.
Should be either the LB or both the nodes.

[defmania]

Hello,

Yes, I've checked all the agents and everything looks fine there, they are pointing at both the nodes.

Apparently the problem was the fact that the appliances have a CNAME, both IP addresses resolve to FQDNs with the same hostname part, but a different domain name. As soon as I went to the VAMI using the FQDN defined on the appliance it worked, the cluster nodes list is populated with all the right nodes.

Thanks,