We're attempting to use an F5 BigIP load-balancer with View and having some problems. Does anyone have any insight or resources that might be of help?
We found a configuration doc from F5 We've configured the load-balancer according to this doc. This is for VDI version 2.x but assuming the requirements are similar.
If we connect to the URL of the load-balancer we consistently get the login page and can authenticate with no problem. However the next page fails to load and says server not found. If we refresh then we get a list of workstations that the account is entitled to. However the page lists the connection status as "not connected" and we can't connected to a virtual desktop. We get an error that the View server is not ready.
Here is what is very odd... If we login to the page the first time we can authenticate but the next page fails to load. If we hit refresh on the browser the View client will download and install. The connection will then work for this first time. But after logging out the subsequent visits fail to work. And If we remove the View Client and reinstall it works for the first visit but fails the next.
In the View Admin console we've disabled SSL connections (the load-balancer will handle SSL). We've also worked with several configuration settings for the View server connection. We've set the 'External URL' to refer to the load-balancer's URL. And also checked or un-checked 'Direct connection to desktop' but no difference.
If we connect directly to the load-balanced name using the View client, we connect with no problems.
If we connect directly to the name/IP of the View Connection server we can connect with no problems (this connection is straight http because SSL is disabled).
Only connecting to the load-balanced https URL gives us a problem.
I think it is a problem with creating the SSL tunneling. My assumption is the initial authentication is through port 443 but once we are on the View server or attempt to RDP to a workstation, then that port needs to be tunneled. But I thought if I enable the 'connect direct to host' and no 'redirect URL' the tunneling wasn't used.
We thought persistence might be a problem. We've reconfigured the load-balancer so only have one server in the pool and disabled persistence but still have the same problem.
What is very bizarre is that the connection works for the first visit and with the client. So don't believe it is a port being blocked. This also makes it a grey area because not sure if it is a VMWare problem or load-balancer.
Thanks for any ideas or suggestions...
