home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!swrinde!sgiblab!rpal.rockwell.com!rpal.rockwell.com!katz
- From: katz@elements.rpal.rockwell.com (Morry Katz)
- Newsgroups: comp.sys.sun.admin
- Subject: Re: Tape Drive Lock-Up
- Date: 25 Jan 93 11:10:12
- Organization: Rockwell International Palo Alto Lab
- Lines: 43
- Message-ID: <KATZ.93Jan25111012@carbon.elements.rpal.rockwell.com>
- References: <1993Jan24.214758.5393@eskimo.com>
- Reply-To: katz@rpal.rockwell.com
- NNTP-Posting-Host: carbon.rpal.rockwell.com
- In-reply-to: nanook@eskimo.com's message of Sun, 24 Jan 1993 21:47:58 GMT
-
- In article <1993Jan24.214758.5393@eskimo.com> nanook@eskimo.com (Robert Dinse) writes:
- >
- >
- > This question pertains to a Sun 3/280 running Sun/OS 4.1.1_U1 and using
- > a 9-track 6250 bpi tape drive.
- >
- > When I use dump to write a disk partition to tape, occasionally I will
- > get the following sequence of errors:
- >
- > xt0: timeout interrupt
- > xtintr: XTE_TIMEOUT reset failed, err= 0x9
- > xt0: hard error bn=0 er=0x40
- > xtwrt_eom:failed in writing EOM 1.
- >
- > Once this happens the drive is hung. Nothing short of a reboot seems to
- > be able to clear it. Everything I try comes back device busy. If I try:
- >
- > mt offline
- >
- > It comes back and says device busy. Powering the drive off and on
- > doesn't fix it either. Once it's hosed, tar, cpio, anything else will say
- > device busy. However the initial error that gets it in this state only
- > occurs when using dump.
- >
- > I have had similiar problems using dump with 8mm 5GB drive on a 4/670
- > (external) and SUN/OS 4.1.3, but power cycling the drive without taking the
- > entire server down seems to fix it.
- >
- > Anybody got any idea what's wrong? Is there a fixed version of dump
- > available that won't hang the drive?
- >
- I believe that there is abug in the SCSI drivers for 4.x. I reported
- this bug to Sun about 1 year ago. They supposedly spent long periods
- of time trying to track down the bug and could never find the source.
- I continue to suspect that the driver contains some state info that is
- not getting reset properly after certain errors are detected.
- --
- ------------------------------------------------------
- Morry Katz
- Rockwell Science Center
- administrator@rpal.rockwell.com (machine administration issues)
- katz@rpal.rockwell.com (other)
- ------------------------------------------------------
-