I did 5 servers. 4/5 worked flawless. I attempted a local upgrade on another machine, it failed. I had to use the CD to complete, which worked fine.
I didn't see any glorious difference. Personally, it's a laterally move, but it does save you from patching all those updates, so I think 3.02 is worth it for NEW installs, however if you are not experiencing problems, read the updates and determine if you really NEED to upgrade.
Other than that, ESX is working perfect as usual, I never really had any problems since 3.01 original CD to say the truth. I only patch to make sure I DON'T have any problems.
ESX 3.0 has NEVER been a problem for us, and the only problems we had were from not reading, ok I say WE, but it's really me...
This board has been a saviour, and since getting more experience with ESX, I have learned much.
ESX is a fantasic product.
See these threads....
3.0.2 upgrade threads - http://www.vmware.com/community/thread.jspa?messageID=712207 and http://www.vmware.com/community/thread.jspa?messageID=712599
Ive upgraded 16 out of my 20 production VM's from 3.0.1 to 3.0.2 be sure to check up on network issues before you upgrade it would seem most upgraders are currently strugling with this without any solution
I upgraded 3 clustered ESX servers managed by VC. No problems here.
After reading other ppls experiences, I made sure the names of the servers were exactly the same in DNS responce, machine name and in the hosts file (there were reported issues involving upper/lower case)
Started by moving the VC DB from "Master" DB to a dedicated DB (the consultant who set it up for us did not do this for some reason). Basically I just made a backup, created a new DB and restored to that one, and changed the ODBC data source. Then I upgraded VC to latest version, rebooted the VC server. All fine.
I then moved all guests away from one of the ESX servers to the two others (live VMotion), upgraded with the .tgz bundle, rebooted. All was fine so i repeated with the two other ones.
I would say everything worked perfectly, no downtime at all for the VMs, and I decided to not disable HA for the cluster before the upgrade (some said that might have been a good idea though).
Maybe worth mentioning I had installed almost all patches except for the very last ones. Because of this I did not have to upgrade the VMWare tools
EDIT: There seems to be some issues after all..
100%. No issues.
VC 2.02 on the other hand...
I am currently just running a test 3.0.2 box, so far no issues.
It appears as if the ping latency problem is the main issue with the upgrade to 3.0.2, so far.
Some people report actual degraded transfer rates (usually with backups) but other folks only have that increase in ping latency without further problems.
Checking the links listed above can provide more information on issues discovered so far.