We have a standard two router two ISP BGP peering arrangement at our edge. About half
of our users connect to their virtual desktop using Horizon client and connecting to the
UAG from various spots on the internet. Two for the second time we had one of the
BGP peers go do. When that happened all of the Horizon connections to the UAG
failed. The screen froze and eventually you were just out.
If you tried again just moment later most of us could get back in There is a side
issue that a number of users go the "Too many active connections" message.
I'll ask about that at the end. But my main question is:
Is there anything that can be done to make Horizon Blast connections to the UAG
more resilient to a partial outage such as I described? Other applications say no
drop at all - our web presence, over a hundred DMVPN spoke sites, vertical
applications, connections to the cloud. No other complaints or anything seen
in monitoring except these Horizon-UAG disconnects. Your thoughts are
appreciated!
Also if you have any thought about avoiding the Too many active connections
message when a bunch of clients get dumped for something like this
that would be interesting too.
Thanks.
