>Number:         4645
>Category:       kern
>Synopsis:       Accessing tape drive causes controller lock-up.
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Dec  5 21:05:02 1997
>Last-Modified:
>Originator:     Dave Burgess
>Organization:
Dave Burgess                   Network Engineer - Nebraska On-Ramp, Inc.
*bsd FAQ Maintainer / SysAdmin for the NetBSD system in my spare bedroom
"Just because something is stupid doesn't mean there isn't someone that 
doesn't want to do it...."
>Release:        1.3_BETA Dec 1 1997
>Environment:
	
System: NetBSD cynjut.neonramp.com 1.3_BETA NetBSD 1.3_BETA (CYNJUT) #1: Mon Dec 1 12:32:08 CST 1997 burgess@cynjut.neonramp.com:/usr/src/sys/arch/i386/compile/CYNJUT i386


>Description:
	When the Archive QIC-525 tape drive is initially scanning the
	tape, any access to the drive causes my Ultrastore 34F
	controller to wedge unrecoverably.  Subsequent reboots will not
	clear the condition:  the error reported is "ccb not allocated".
	I noticed this earlier in the week and ascribed it to a bad
	cable.  I have since replaced the cable.  All of the other SCSI
	problems went away except for this one.  
>How-To-Repeat:
	Turn on the computer with a tape in the tape drive.  Escape out
	of the memory test and hit return on the boot selector.  The
	tape is still scanning for format at this point.  As soon as the
	Ultrastore controller scans the tape drive (which is ID-6), the
	machine will dump core to the IDE drive and reboot.  The
	subsequent reboot will have the Ultrastore controller either
	completely unrecognized (as in 'disappeared') or will error out
	at the first drive probe.  Another way to cause this is to put a
	tape in the tape drive and try and access it before the drive
	has settled.  I've even had it happen once when I Put in a CD
	and tried to access it before the drive was completely spun up.
	That was before I put in the new cable, though, so that may be
	an abheration.
>Fix:
	The only work-around I've found it "don't do that".  Let the
	tape drives and other SCSI devices settle before attacking them.
>Audit-Trail:
>Unformatted: