So, connect to trophy. Then telnet to whichever array has a flashy light (t3a or t3b). root passwords are in the standard place.

Once logged in to the array, the command fru stat will show you which disk is bad. Make sure that this makes sense, given the amber light that ought to be flashing on the front of the array.

Here's a rough transcript of one of my conversations with Sun support regarding a bad disk (14 July 2004, take it for what it's worth given its age).

Call 1-800 number (800-722-4786). Press 1 for English, then 2 for new support request, then 2 for warranty, then 1 for warranty. I pressed # to say I didn't know when it was purchased.

I spoke to Cara/Tara (didn't quite catch the name). I said I had no record of the serial number, she asked if it was Canada or US. Told her Waterloo, sweet, that got it. I told her it's connected to a Sunfire 3800 running Solaris 8. She made me the contact for this request, confirmed my address, and gave me a case ID. She had me do the "fru stat" thing to show which disk was bad. That was sufficient, and she sent me off to Sharon in the Storage/SAN group, who shipped me the disk same-day. After that I didn't do the paperwork, I cheated and gave it all to Tom Gal who happened to be on campus looking at... a bad disk... in IST. Sun had asked if I wanted somebody to come do the disk replacement, I said no but perhaps should have said yes.

In total I spent about 30 minutes on the phone. Switching the disk out took about 15 minutes and another several hours to rebuild the RAID (but I didn't have to watch it). proc list shows what things the array is doing. It ought to look something like this:

t3a:    /:<2>proc list

VOLUME         CMD_REF PERCENT    TIME COMMAND
v0               22592      48   28:51 vol recon
t3a:/:<3>
</code>

-- Main.MikePatterson - 10 November 2004

And again on the 11th of November 2004, the procedure was very similar.

-----

Serial Numbers

t3a - 113H4781
t3b - 113H4773