Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!usc!apple!snorkelwacker!ira.uka.de!fauern!tub!gmdtub!tmh
From: tmh@prosun.first.gmd.de (Thomas Hoberg)
Newsgroups: comp.unix.i386
Subject: Re: Tape cartridge hung
Message-ID: <278@prosun.first.gmd.de>
Date: 13 Aug 90 17:54:05 GMT
References: <1990Aug2.074751.4386@oct1.UUCP> <8@utekza.UUCP>
Organization: GMD-FIRST, D-1000 Berlin 12, Germany (West)
Lines: 12

I have an Archive Viper QIC02 60MB streamer with an Archive controller
that will hang, too, sometimes. I am under the impression though, that
this will only occur when a couple of tape accesses come within a short
time (e.g. rewind, status, cartridge change, rewind, cpio -itvc, *BREAK*,
cpio ivxdmc ....). If I leave it alone for a minute after three operations
or so, it won't hang. Whenever it hangs, there is no way to kill the
offending process: it seems to be blocked on a read in the kernel and
no amount of signalling will wake it up. This is especially nasty when
the offending command was run on the console, because uadmin 2 0, init 0 or 
shutdown can only be executed from the console (going single user takes
care of that), still that leaves a file system busy. I have had this 
bug consistently with ISC 1.06, 2.0.0 and 2.0.2.