home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!cs.utexas.edu!sun-barr!ames!sgi!fido!zola!zuni!anchor!olson
- From: olson@anchor.esd.sgi.com (Dave Olson)
- Subject: Re: More Exabyte 8500c (and SCSI) questions/woes
- Message-ID: <sj448gk@zuni.esd.sgi.com>
- Sender: news@zuni.esd.sgi.com (Net News)
- Organization: Silicon Graphics, Inc. Mountain View, CA
- References: <1992Nov19.200617.27850@ncsu.edu>
- Date: Fri, 20 Nov 92 03:25:04 GMT
- Lines: 48
-
- In <1992Nov19.200617.27850@ncsu.edu> garvin@cfd.mae.ncsu.edu (Michael Garvin) writes:
- | The minor numbers were upped 16 from the default 224-227 to match the 8500c
- | (0x8c in the tpsc config file) entry. I then tried tarring to the drive, which
- | died with the following error:
- |
- | tar: tape write error: I/O error
- |
- | and produced the following console output (note the tape drive hiccuped first):
- |
- | WARNING: dma_map: address not word aligned
- | sc0,7,0: cmd=0xA disconnected on non-word boundary (addr=C02185A9,
- | 0x0 left). Resetting SCSI bus
- | dks0d1s6: retrying request
- | dks0d2s6: retrying request
-
- All SCSI DMA must be word (32 bit) aligned. A modeselect is supposed
- to force this for Exabyte, but has never seemed to be reliable. On the
- 8200, you could set it with a dip switch on the MX board. On the 8500,
- it has to be set through the monitor port, or by loading a new firmware
- tape. You'll have to go to your distributor or Exabyte on this one.
-
- | Our setup is a Power Series 4D/440. SCSI controller 0 has an internal 1.6 gig
- | system disk (ID 1), an Iris File on the rear external connector with four 2 gig
- | drives (IDs 2, 3, 4, and 5), an internal CD-ROM (ID 6), and the 8500c mounted
- | internally (ID 7). We've also got a Jaguar differential SCSI controller with
- | four drives on each channel.
- |
- | My questions are: am I going too far with that SCSI chain? I have a
- | sinking feeling that I'm pushing the length limit, giving me the error I saw
- | (although the 8500c failing might give me that). Are my entries for the 8500c
- | in the tpsc config file correct? Have I correctly made the devices (in
- | particular, did I get the minor number correct and do they correspond to the
- | 8500c in compression mode)? And finally, would someone have the correct density
- | to feed to BSD dump for this drive (I calculated something around 47.5 million,
- | which sounds wrong). Thanks for any help, I'm banging my head at this point.
-
- You could be pushing the max cable length, but I don't think that has
- anything to do with your problem.
-
- On dump, the numbers are completely bogus anyway. Just use
- something like 25000 and 54000, assuming an average of 50% compression.
- One of the problems with compression, of course, is that you never know
- just what you are going to get, and our dump doesn't yet handle EOT,
- unlike Ultrix; that was one thing they got right.
- --
- Let no one tell me that silence gives consent, | Dave Olson
- because whoever is silent dissents. | Silicon Graphics, Inc.
- Maria Isabel Barreno | olson@sgi.com
-