We've analyzed the process of disabling Auto-Deployment described in our KB article and we think it lacks a note about manual re-deployment of collectors. You should wait another hour or perform another health service restart to be able to re-deploy collector manually, because collector auto-deployment workflows are still there and they will try to remove/uninstall a collector you are trying to install on the management server. Here is the process:
- SCOM Workflow on VES discovers that servers with auto-deployed collectors are outside of the resource pool (because resource pool name in VES is empty) and collectors should be uninstalled. This could take up to 5 hours by default and two health service resets.
- Collector installer workflow uninstalls the collector, could take another hour or a health service restart on this management server.
- SCOM Workflow on VES detects that collector installer removed the collector and can be undiscovered, Also could take an hour and another health service restart
We think that you probably tried to install Collector after step 2, but before step 3 happened. In this case Collector installer workflow will try to uninstall the collector again, producing another orphaned collector (collector will be orphaned after uninstall and you have to unregister it in the UI).
To make sure collector auto-deployment is not doing anything, could you try to use Discovered Inventory view to display "Veeam VMware Collector Installer" object? The view should be empty if all installers are undiscovered, if it's not empty there could be autodeployment workflows still trying to install/uninstall collectors. Could you send us a screenshot with all properties if there is still something in this view?