I have had a bit of a search for clarification on this, but have only found info on draining session hosts.
I need to make some amendments to our 2012 R2 gateway server VMs, but due to the nature of our business, the gateway servers are in use most hours of the day.
My plan is to put one of the gateway servers into drainstop mode through Microsoft NLB, but have concerns that currently connected users may see some disconnects when the active connection refreshes.
Can anyone confirm that if I put the server intro drainstop mode, current connections will remain unaffected until the user disconnects from their end?
Seems that the answer to this is yes, users are affected, even though they shouldn't be.
Rather than behaving as I expected and allowing the connection to continue until the connection is dropped (and then directing new connections to other nodes), enabling drainstop on a host does a "converge", whereby the connection is dropped and brought up on another node.
I was doing this from home, and as soon as I drain stopped the node I was connected to, I was temporarily disconnected.
According to KB Article, this should not happen. This may mean our NLB cluster is not configured correctly?