davy@pur-ee.UUCP (Dave Curry) (10/17/85)
We've been running VMS on our UNIX disks for a few years now. We don't run a full-blown lots-of-disks version; just something for the hardware guys to run diagnostics, patch microcode, etc. with. George Goble did the work on it, I asked him about it and below is roughly what goes on. We have a special file system on each machine called the reload area. Basically, it's a 65MB (131,000 sectors (cylinders?) on an Eagle) scratch area for loading in tapes, etc. The file system is made to start at the "front" (first partition) on the disk. I should probably note here that we partition our disks differently than the way Berkeley does. Anyway, to make VMS work on the CDC9766 (like an RM05) drives, George went in and adb'd the VMS binary. He changed the table for the disk drive to convince VMS that the drive was only N sectors (cylinders?) [fill in the number of sectors/cylinders in 65MB on a 9766 for N] long. Then he hacked up the various bootstrap things, and so on. Dump the whole mess off to tape, and you're done. The hardware guys just "dd" the tape onto the reload area, and go. For Eagles, George had to hack in a new disk drive. An eagle is an RM02, but VMS doesn't grok those, so George made it an RM03 (or the other way around). After that, the procedure is the same. It seems to work. I'm afraid the description above is probably somewhat skimpy... I don't know too much about this sort of thing. You could probably mail George (pur-ee!ghg, ghg@purdue-ecn) if you really need to do something like this... --Dave Curry