Adding HA to an environment does seem to make administration easier than in a non-HA environment. There are very few changes you can make that result in needing to to reboot the entire cluster of machines at once. Most of the time you can play with one machine at a time.
Some services, however, still have the same issues as standalone. For example, CIFS shares. Regardless of if you're offering high availability, you still generate disconnects when the original host goes down.
With HA you have to make sure you've upgraded your infrastructure not only at points A and B, but with the networking in between. It makes no sense to support failing over a datastore when a single port on your switch can fail and take it all down.
Some services, however, still have the same issues as standalone. For example, CIFS shares. Regardless of if you're offering high availability, you still generate disconnects when the original host goes down.
With HA you have to make sure you've upgraded your infrastructure not only at points A and B, but with the networking in between. It makes no sense to support failing over a datastore when a single port on your switch can fail and take it all down.