i'm testing View 4.6 for internal use and today my 3 users reported that they cannot connect to their desktops.
The status reported for the desktops in the View Manager is "Available (missing)".
I've noticed you can still connect with RDP if you point directly to the IP of the Virtual Desktop, shows as "Connected (missing)"
I've noticed it does not show the host anymore nor the datastores where the Desktops are located, seems like a communication problem with vCenter.
I've rebooted both vCenter and the View Connection Server with no results. View is 4.6 and vCenter is 4.1 just like all the hosts.
I'm attaching a screenshot to show you what i'm talking about.
If you create a new pool are you able to browse the host and datastores located in vCenter? Have you taken a look at the logs to see if you can determine where the breakdown in communication is happening?
UPDATE: Seems like i can't browse the datastore from the view manager.
I removed all pools, removed vCenter from View Manager, readded vCenter and when trying to add a manual pool pointing to a vDesktop i was using before it shows that there is not virtual machine (compatible or incompatible).
Maybe this will help you guys to help me 🙂
What happens if you take the account that you are using to integrate vCenter and simply login using the vSphere Client? Are you able to browse all the host, datastores, etc?
You'll only see the (missing) state if View was able to talk to vCenter and it was told the VM object no longer exists in the inventory. Sounds like you've reinstalled vCenter or removed the ESX host from it, therefore losing all the relevant objects. If you connect to vCenter directly using the vSphere client can you still see the VMs? Rebuilding the inventory in vCenter should allow View to find the missing VMs again, references to datastores and templates may need to be repaired by editing the pool settings afterwards.
I've been told yesterday the resource pools have been replaced with vApps because they wanted to leverage the start order feature.
I checked and all permissions have been retained even from the datastores.
Maybe View can't be put in a vApp? Do you think restoring the previous configuration could help?
Sounds like the reason it's not working anymore...
Do you have a procedure to rebuild the index in vCenter inventory anyway? Sounds it might be helpful to know it.
I'll start reconfiguring right now putting back the Resource Pools and see what happens. I will report ASAP.
Sorry, what I meant by rebuilding was just putting the objects back into the same folder structure as before rather than anything fancy. View will find VMs that have moved but there are some objects, such as resource pools and clusters, which are referenced by path and therefore need to be recreated in the original location or View's pool settings need to be updated to the new name.
I can confirm the vApps were the problem.
I reconfigured the resourced pools previously configured, rebooted the view manager and now it's working again.
I awarded the points to Mike Pryor and i thank him for his support as much as all the others who helped.
Last thing i would like to know is if i can put vCenter in a vApp leaving all the other View machines in a resource pool or if this will still cause problems.
Since View doesn't manage the VMs running vCenter or the connection server instances themselves, they're essentially equivalent to physical servers from View's perspective so you could theoretically put them in vApps or anything else that takes your fancy (not that I've bothered to do this myself, and of course assuming that you're not changing guest settings like hostname, network location etc.).