Yes, you could stick to the setup that you explained. The only downside is that there is a manual action required for the cutover to the DR site.
What you could use is CPA with the homesite assignment to provide the automatic failover.
It will ensure an automatic failover to the other side when the resources (desktop or applications) are unavailable.
More info on the homesite can be found here: Using Home Sites
Another solution would be to use for example AWS Route 53 for the GSLB functionality.
Where you can have a active-passive failover setup for the primary and DR site.
Andrew Morgan did write a nice blog on the VMware blogs site on setting up this: Global Load Balancing with VMware Horizon and Amazon Route 53
The cost for running this is really low, I thought it was less than 15 dollars per month when I did the calculation for a Customer.
Hope this helps!
I'd still use it and just give the users the dr address when needed. That way once production is back up users when they log back in will get redirected to the home site when it ready. Ever way you should also look at storage groups to replicate the appstacks, as importing them manually is annoying.
Things to consider:
User data / File Share - How would you setup reference to this in a DR Site? You do separate OU with a seperate GPO for pointing your UEM config/profiles? This requires a little more management but is probably your only method.
App Volumes - You can use the method of putting the appvolume manager to the registry entries from this VMware Knowledge Base , but I have found this to not go in order and select the first to respond to requests or random, and at times it would ignore one. Alternative would be to modify the image when you replicate it to your DR site for this one simple registry change, I had success with this.
Connection Servers - Honestly, although I use Cloud Pods, when I did not have a GSLB I preferred pushing the connection server string to my users. Today I rely on GSLB but what you can do is create a DNS entry, point it to your prod but for your cert include the alternative names.
Yes so if i was to use CPA i was going to set everyones home site to prod obviously but how would this work in terms of adding the DR connection servers into the mix? Without a loadbalancer at the secondary site i'd need to add them to the primary sites loadbalancer so i can use the same DNS name / VIP? Downside of that is im tunnelling connections via https in Horizon so could that mean connections could potentially be tunnelled via my DR site to the user homesite if the loadbalancer was to throw them at the DR connection servers?
how does CPA know if resources are unavailable? I would have to add my DR connections servers to the prod loabalancer at site 1 wouldn't i to keep using the same DNS / VIP? I dont have a LB at site 2. Or could i just leave my connection servers standalone at site 2 with no LB and add them to CPA and if site 1 is down it will automagically point people to site 2?
You need to point your users to a connection server, if you point them to connection servers at site 1, and site 1 is down, they have no idea where to connect.
CPA is not a load balancer for your connection servers, it utilizes your other sites as additional capacity. You need a GSLB for load balancing your connection servers across multiple sites. If you don't have one, you would need to utilize DNS for this, which is not ideal.