\_ SMTP quoth George Toft on 3/22/2001 23:18 as having spake thusly: \_ \_ I have a Sparc2 with Red Hat 5.2 installed. The hard drive is correctly \_ identified if I type "probe-scsi" at the prom ok prompt, but it hangs \_ when I try to boot from the drive. \_ \_ I suspect a bad drive. It spins up, prom identifies it, but the Red \_ Hat installer won't read from it. I've set the jumper for termination, \_ and that did nothing. Is it scsi id 0? At least in x86 land, your boot drive needs to be zero. Unless that's changed in the real world...it hasn't changed in mine. :-) If you have another scsi chain handy, plug the drive into it and see what kind of mileage you get. I once recovered data off of NTFS this way. The box cratered (suprise), I chained the drive into my dual boot...NT wedged (suprise [1]) so I said, wtf, let's give the dev kernel w/ NTFS support a whirl..."would you like fries with that data?" They let me crow about how my free replacement was better for about a week vefore they told me they had had enough of it. David [1] This is probably on account of NT being stupid and auto labeling its drives for you in a bass ackwards way: it goes breadth first across the "hard" disk drives, then comes back for more. eg: one drive, the null span: c: hda1, d: hda2, e: hda3 ... x: cdrom eg: two drives: c: hda1, d: hdb1, e: hda2, f: hdb2, ... x: cdrom eg: three drives: c: hda1, d: hdb1, e: hdc1, .... x: cdrom So what's the problem, I hear you ask? What if you don't install your OS on C:? I had it on D:, with apps/transfer space on C:. With the second drive in there... D: wasn't hda2 any more it was hdb1...no OS found, *poof* bluescreen. I actually figured this out sometime later as I was installing NT on a different box in another company.... a box with a zip drive. Install didn't see it, and I did the OS on D: again. Service pack install to latest and greatest saw it...and added it as hdb, of "hard" disk nature (as opposed to cd drives which are ignored for this). Now all of a sudden, no boot after service pack install, also because D: wasn't pointing to hda2 anymore. Brainchild idea that spanning idea was. *sigh*