Recently in my lab, I was frustrated over an attempted upgrade to version 7.3 from version 7.2 of vRealize Automation. It had been a while since I had last used my vRA deployment, so I naturally assumed I was probably bound for some sort of issue due to neglect. However, typically I can figure these things out pretty quickly and go along my way. This time was different and the solution, was frankly disappointing. Perhaps a lack of understanding on my part, but I think it was more of a lack of documentation from VMware on vRealize Automation.
When I attempted to upgrade the vRA appliance to 7.3, I received an error in the Upgrade panel stating the following:
The VMware certificate was not installed on the nodes listed below. Make sure the Management Agent is up and running on them and that it has connectivity to the VA.
It then displayed a single IaaS server from my lab. When I took a look at the general vRA Setting – Host Settings, I was baffled to see that I had two IaaS servers with the same name, only spelled different through capitalization. Considering this is a Linux appliance, two totally unique names for the environment.
At some point in my lab, I must have rebuilt the IaaS server, but had changed the name from a more complex upper/lower case one to a simple lower case one. The answer is pretty straightforward then. Delete the obsolete IaaS server. And this is where the fun began.
No place in the documentation discusses how to remove an IaaS server from the appliance database. Assuming this was vRA, I was beginning to think it would be some sort of command line action to remove the entry. However, nothing could be found anywhere on how to remove it on the internet. I checked out some various commands on the appliance and I was able to find ways to see a list of the IaaS servers through the “vra-command list-nodes” command, but the vra-command did not have any options for removing nodes.
Even looking for “IaaS uninstall” didn’t lead to any results of value and I even went so far as to post on the VMware Communities for an answer, for which I received a single response that, while interesting, was definitely not going to resolve the issue for me.
Finally, I landed on the answer after what should not have taken nearly as long as it should have. The answer was in the vRA appliance management under the Cluster settings under vRA Settings.
Nicely tucked down at the bottom, a list of the nodes in the environment…and a Delete button! That kind of information would have been helpful in some sort of documentation somewhere, perhaps titled “How to delete a node” or “How to remove IaaS from vRA”. Clearly, it had been a while since I switched my IaaS servers around considering almost a year had passed since the old name had checked in. I hit the Delete button and the node was removed a minute later.
Going back to the upgrade, it now worked properly and I was able to get to 7.3.
A few things to note. I took snapshots before removing the node, just like any other action in vRA which may cause destruction. It’s always a good practice. Additionally, I confirmed on the Host Settings page that the vRA appliance now reported the IaaS properly and having a good certificate. Lastly, considering the Cluster page also shows version information for your vRA environment, it’s a good go-to for determining the version status of your environment.
Thanks and I hope that helps.