As part of my day to day job I get to experience some weird and wonderful issues, some product related and some customer. In this particular case it was a bit of both.
I was allocated several days to perform an in-place point release upgrade of an existing customer platform which on the face of it is not a large task as long as the appropriate care is taken conducting the pre-flight checks before you start hitting upgrade buttons.
I spent a full day performing checks such as verifying the platform could be shut down and started back up from cold and still be functional, verifying no errors in log files and many others (the full list of what you should be checking is in the online documentation).
Towards the end of the day we got downloading the update. Top tip, download the ISO file for the update rather than downloading the update live. It is significantly faster, particularly if you have to re-run the update due to issues.
Once we had appropriate snapshots, SQL backups etc I continued with the upgrade and straight away had a pre-flight check failure from the vRA appliance itself. Here’s a screenshot of the failure.
After much digging around and verifying the IIS configuration on both web nodes I found that both web sites had been renamed from “Default Web Site” to something else. The pre-flight check of the upgrade package was hard coded to look for only “Default Web Site” and if it couldn’t find it then the check would fail.
To get around this issue, the web sites on both nodes had to be renamed back to “Default Web Site” to allow the pre-flight checks to pass before being renamed back again to their original name after the pre-flight checks had completed, to allow the upgrade to work. The web nodes had to be restarted after each rename to allow IIS to function.
Once all these changes were completed the upgrade sailed through and completed in just over an hour for all components.