goss@SNOW-WHITE.MERIT-TECH.COM (Mike Goss) (03/21/90)
Anyone out there tried to hook up a Hewlett-Packard SCSI disk (in particular, model # HP 97548-S) to an SGI machine (especially a Power Series or PI)? This HP disk doesn't appear to be in the table of supported disks in "fx", so it looks risky unless someone has already been able to get it to work. I'd appreciate any reports of success (or failure). Thanks. ------------------------------ Mike Goss Merit Technology Inc. (214)733-7018 goss@snow-white.merit-tech.com
olson@anchor.sgi.com (Dave Olson) (03/22/90)
In article <9003201748.AA13587@snow-white.merit-tech.com> goss@SNOW-WHITE.MERIT-TECH.COM (Mike Goss) writes: | Anyone out there tried to hook up a Hewlett-Packard SCSI disk (in | particular, model # HP 97548-S) to an SGI machine (especially a Power | Series or PI)? This HP disk doesn't appear to be in the table of | supported disks in "fx", so it looks risky unless someone has already | been able to get it to work. I'd appreciate any reports of success (or | failure). Thanks. There is no internal table of scsi disks in fx (there is for the other drive types SGI supports). fx uses the mode sense command to determine the required drive information. If the HP drive follows the 'de facto' mode sense pages and layout reasonably closely, then it may work. I've never worked with this drive, so I can't guess. As time goes on, we try to make the scsi portion of fx less dependant on embedded info about the page sizes, but we are pretty much stuck with the info layout, and will be to some extent even when scsi 2 is implemented by more drives. If fx is able to determine the info it thinks it needs, then it will automatically create (in memory) a volume header for the drive. If you do a /label/show/all at this point, and the values look 'reasonable', then write out the volume header and try exercising the disk. If that works, you are probably in pretty good shape for using the disk, although I'd recommend running 20-30 sequential wr-cmp exercise passes under unix (not the standalone version), followed by a fair number of random and butterfly tests, followed by a few more sequential, before I tried making a filesystem on it. Dave Olson Life would be so much easier if we could just look at the source code.
mjb%hoosier.utah.edu@cs.utah.edu (Mark Bradakis) (03/29/90)
In article <9003201748.AA13587@snow-white.merit-tech.com> goss@SNOW-WHITE.MERIT-TECH.COM (Mike Goss) writes: >Anyone out there tried to hook up a Hewlett-Packard SCSI disk (in >particular, model # HP 97548-S) to an SGI machine (especially a Power >Series or PI)? We tried it, it didn't work. It seems there is some disagreement between SGI's implementation of SCSI and what the HP disk assumes. Plus SGI, like many other vendors, will not spend a lot of time trying to get someone else's stuff working on their machines. mjb. mjb@hoosier.utah.edu "Tell the truth, explain to me how you got this need for speed."