Fixed by removing the entries from the dbo.VMssVMS table within the Veeam Database.
I believe these are related to a mounted data PowerNFS data store? No VM's were mounted but the data store was still active.
In either case, in future it would be nice if the installer checked these things if they are known gotchas, especially as once you have started the install you can't exactly launch the console to disable it.
No, active vPowerNFS datastore has nothing to deal with this issue. The dbo.VMssVMS table holds the list of active Instant VM Recovery processes, as long as there is some content in this table, the setup program will not be able to modify the table as required (see 2nd bullet under common upgrade issues in the sticky Known Issues topic). Thanks!
Also saw this issue today. Seems that an Instant Recovery session did not clear down from the database, although it was no longer present according to the GUI. We checked this specifically before starting the upgrade. We could see the problematic row in the table, so backed up the DB and removed with :