Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!ucsd!pacbell.com!decwrl!sgi!shinobu!odin!anchor!olson From: olson@anchor.esd.sgi.com (Dave Olson) Newsgroups: comp.sys.sgi Subject: Re: SCSI timeouts Keywords: SCSI Message-ID: <10819@odin.corp.sgi.com> Date: 24 Jul 90 20:43:08 GMT References: <822@ki.UUCP> Sender: news@odin.corp.sgi.com Organization: Silicon Graphics, Inc. Mountain View, CA Lines: 41 In <822@ki.UUCP> dwatts@ki.UUCP (Dan Watts) writes: | Lately I've been getting a few errors like | | sc0,2: Resetting SCSI bus: timeout after 60 sec | | These seem to occur randomly and infrequently. My setup is | a Personal Iris with QIC-24, QIC-150, internal 350MB+ drive | and an external 650MB+ drive. I don't recall seeeing this | error before I put the external drive on. All the hardware | works fine otherwise. I was able to get the error to repeat | itself today by trying to tar some files from the external | drive to a 3M DC615A tape in the QIC-150 drive. Should I | put my external disk at a lower SCSI address than the tape | drives? Some info about what OS release you are running, and the drive types (mt -t /dev/.... status for the tape drives, and the drive name that fx 'dksc(0,#)' gives you for the drives) might help. The 60 second timeout probably indicates that the tape drive was not able to reconnect after accepting a command for some reason. This could be as simple as the drive taking far longer for some operation than the qualified drives. It could also be due to some problem with one of the disk drives. Finally, it could be a cabling or termination problem. Check to be sure that you haven't configured anything in a 'star' configuration (you need a straight through connection from the system to the final drive, with only short stubs coming off of the bus). Finally, check to be sure that there are NO terminators on any device, except possibly the last external device on the bus, if it doesn't have external termination. Changing the ID of any of the devices is very unlikely to have any effect on the situation, unless you have ID'ed 2 of them the same, which is unlikely from your hinv output. -- Dave Olson Life would be so much easier if we could just look at the source code.