This issue caught me out a few days ago. Having designed and coded automation and orchestration services for one of the UK’s largest service providers for the last year I was somewhat surprised to see many of my deployed NSX test objects (NSX virtual servers, logical networks etc.) become unusable overnight from within the vRA 7.2 portal.
By unusable, I mean when you click on them none of their details show and when you try and run any day 2 actions against them they all fail with application errors on the portal. Unfortunately there is no easy fix for this without hacking the postgres database but the cause is quite simple to explain.
The NSX objects that are within the vRA portal are linked via internal identifiers to the NSX plugin configured on the vRO server. If you re-configure the NSX plugin within vRO (i.e. remove the NSX connection and re-create it) then the NSX objects within vRA become dis-associated from the objects within NSX because the NSX ID no longer matches up.
In my case somebody had decided to try and re-add the NSX connections in vRO with an account other than the NSX admin account. When they had worked out they couldn’t do that they had left the environment configured back with the starting configuration however by that time the damage had already been done.
I suspect that the damage could be repaired by manually updating postgres records however as this was a development platform and I was in a hurry, I simply removed the objects from the postgres database and then from NSX before deploying some new ones.