How to configure real High Availability with automatic reconnect?

[tomcek]

Hello community,

I have not yet figured out how to configure a real HA. What I mean with real:

During a session, when one of our UAGs fails (for what ever reason) the session for the users is interrupted and they see the typical warning about connection lost and the time starts to count down.

Now I would expect the second UAG, which becomes the primary node at that moment, to resume the connection for the users. Unfortunately, this doesn't happen. The seconds in the Horizon client count down and no automatic reconnect happens. Instead, the users have to reconnect manually. Can this be automated like I would expect as a real HA?

 

P.S. Nice to be back in the new Omnissa community. Great that all the old threads have been migrated. 👍

[Graeme Gordon]

There is no way for an active user Horizon session to fail-over automatically to another UAG appliance.

The users protocol session (which is the ongoing session that gets interrupted if a UAG fails) has to go through the same UAG that the user authenticated through. We cannot hand off the protocol session to another UAG  as it would not understand who the user was and the state of their authentication.

The correct action is for the user to reconnect.

See this guide for an explanation of how the Horizon connection works.

https://techzone.omnissa.com/resource/understand-and-troubleshoot-horizon-connections#external-connections

[tomcek]

Thank you, Graeme, for the explanation. That helps to understand.