I recently set up an environment with two vCenter Servers using linked mode. The first vCenter Server 5 had been an existing vSphere install that had been upgraded several times and the second vCenter Server was a fresh install of vSphere 5 at a second site. Some time after this was completed, while connected to the first vCenter Server via the vSphere Client, I noticed that several of the larger datstores on hosts managed by the new vCenter Server at the secondary site had alarms triggered for datastore usage on disk. I eventually realized that the default values were still in use on the new vCenter Server, and this second environment regularly exceeds these defaults. I went to edit the alarm settings of the "Datastore usage on disk" alarm definition and clicked the Triggers tab. I was presented with the following message: The triggers for this alarm cannot be displayed or changed through the vSphere Client. Use the vSphere API to modify these alarm triggers.
My first thought was that this must be due to the clean install of vCenter Server 5 at the second site and that there was no way I was going to have to resort to using APIs to make this change. A quick Google search didn't turn up much in the way of results, and then I had a thought. What if linked mode was somehow behind this? I started up another instance of the vSphere Client and connected directly to the vCenter Server at the secondary site. Just as expected, everything works normally with this approach and the trigger values can be adjusted as expected. I made the changes and the warnings cleared on these datastores!
I tried several other alarm definitions and this behavior is not consistent. Some can be modified across vCenter Servers and others present the same message documented above. Hopefully this information will help anyone else that stumbles upon this condition.
Thanks for reading,
Brian
