Bill Bradford wrote: In the UK, in 2000, they wouldn't ship a power supply without an attendant engineer to perform the complex task of undoing the thumbscrews. (With the attendant cost, as well - of course...) Whether this was global policy or just a local thing, or just SGI being bloody minded for the sake of it I wouldn't like to comment :-). Of course, when said engineer did come round, he cheered me up immensely with "Oh, yes, this model of PSU was known to be bad, they always blow; the replacement should be fine." Well great, thanks for telling me when I needed to know... I also forgot to mention the sheer awfulness of IRIX in my previous list of gripes... Anyone else had to deal with a corrupt XFS filesystem on a production machine? "Excuse me, my XFS filesystem is corrupt and I need it fixed," "well, XFS can't get corrupted so we don't have any tools to repair it." "OK, lets pretend XFS isn't corrupt but that IRIX panics every time it reads a particular directory entry, how do I repair it?" "Sorry, XFS cannot be corrupt, so there is no way to repair it..." Cue a long night of restoring from backups :-(. Fortunately for Sun, there ain't any SGI systems still in production use round my end of the world any more... Cheers, Tim. -- Tim Walls at home in Croydon - Reply to tim(a)snowgoons.fsnet.co.uk