> My original post was:
> I think I know what the answer is, but just by chance maybe some else has
> figured out a work around. Here's the problem: I have a process that is
> hung (ufsdump) and is now owned by INIT (pid=1). I cannot kill the
> process and a kill -HUP 1 doesn't clear it either.
>
> Is there a way to terminate the process short of rebooting the machine,
> which is not an option anytime this month. Not being able to backup
> the system is not an option either. Ah yes, the ol' rock and a hard
> place.
Many thanks to:
wbmoore@fedex.com
jayl@latticesemi.com
brion@dia.stata.ma.us
amonf@SubaruSparcDev.subaru1.com
bobw@filmworks.com
bchievers@karoshi.mitretek.org
Glenn.Satchell@Uniq.com.au
gibian@stars1.hanscom.af.mil
and any others I may have left off (sorry).
The "Golden Sun" award goes to Walter Moore (wbmoore@fedex.com) for a most
innovative solution. Walter's suggestion, assuming that kill -9 doesn't
work, was to change the SCSI ID on the tape drive and run drvconfig and
tapes. He suggested trying this with the unit turned off and doing a
kill -9, but the process still didn't die. Chaning the SCSI ID does work
and I was able to get the backup processes going again.
Jay Lessert (jayl@latticesemi.com) suggested running truss -p on the ufsdump
pid. I did and this confirmed that the process was completely defunct and
was not responding to any signals.
Many thanks to all.
Ira
. .
\\ //
\\^ ^//
( o o )
oOOo __O__ oOOo
********************************************************************************
* Ira Childress * IBM Global Services *
* Systems Manager Integrator * *
* IGS/Comdata * *
* Off 615/370-7662 ********************************************
* FAX 615/371-3167 * *
* irac@comdata.com * Comdata Corporation *
* Ira.Childress@nashville.com * Transportation/Finance/Gaming *
********************************************************************************
* There are two major products to come out of Berkeley: LSD and UNIX. *
* We do not believe this to be a coincidence. *
********************************************************************************
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:56 CDT