<updated> Why external array failure does not cause cluster failover

For clarification - the external array is an HP MSA30 that holds the Data for our Domino server, both power supplies died (so much for redundancy when the second fails while the first is being swapped out).

Domino 6.5.5 FP2, on Windows 2003. An external array failed, causing windows to report errors updating $MFT, application pop-ups from Domino about losing data updating various .NSF files, but yet no failover occurred until we shut down the server (we couldn’t get domino to shut down for all the errors).

Has anyone else seen this? (I searched and didn’t find anything) Have you reported it to IBM? What did they say? Why isn’t Domino smart enough to at least say that it’s “BUSY” if it can’t write to the hard drive?

The problem this caused is that for cached design elements (memo form in mailbox, etc) users could still work. So, they could compose and send messages (I have logs of e-mail being sent during the time the server was running without the array) but they weren’t able to open mail (that was not still in the server’s document cache) etc. VERY annoying.

I know I need to file a report…just wanted to hear other experiences.