A recent client experience reminded me that advising clients about old copies of virtual machines is critical. I normally say to users on vmware and hyperV to clearly rename the old copies that are made of virtual machines to prevent them getting confused with a live version. However thats exactly what led to the clients predicament when due to a comms fault access was lost to a vm running on a hyperv host. Unfortunately the user dispatched to deal with the problem was unaware that the vm in question was not on the host where they found a shut down VM – but this old vm was restarted anyway.
Suffice to say all sorts of strange behaviours including trust relationship errors trying to connect to the vm with rdp led the client on a merry dance and the issue was finally resolved when they realised that ping responses were coming back from the vm – even though they were convinced it was shut down.
The trust issue was resolved by dropping the vm off the domain and re-adding it.
So the lesson here is – document your vms and if there are copies make sure you rename them – perhaps with the date that they were made or shut down