[5087] in testers

home help back first fref pref prev next nref lref last post

Re: sun update problem

daemon@ATHENA.MIT.EDU (Garry Zacheiss)
Thu Mar 21 07:29:37 2002

Message-Id: <200203211229.HAA11143@never-never-land.mit.edu>
To: Mitchell E Berger <mitchb@MIT.EDU>
cc: Jonathon Weiss <jweiss@MIT.EDU>, Greg Hudson <ghudson@MIT.EDU>,
        Garry Zacheiss <zacheiss@MIT.EDU>, testers@MIT.EDU
In-Reply-To: Your message of "Thu, 21 Mar 2002 04:38:46 EST."
             <200203210938.EAA13112@byte-me.mit.edu> 
Date: Thu, 21 Mar 2002 07:29:33 -0500
From: Garry Zacheiss <zacheiss@MIT.EDU>

>> Both byte-me and flying-shards-of-metal did the same thing.  I looked at
>> the patch, and it's a ginormous accumulation of things.  Reading its
>> prepatch script
>> (/afs/dev/system/sun4x_58/install/patches/patches.link/108528-13/prepatch)
>> indicates that for some reason (a bugid is associated with it), it needs
>> to kill powerd before installing, and tries to do so.  I imagine this is
>> what we're seeing here.

   Nope.  powerd isn't running and isn't getting killed; the "Killed"
message is actually from from the "kill -9 $child_pid", which is killing
the subshell that's doing:

( sleep $SLEEP_TIME ; kill -ALRM $$ ) &

to time out the question that's being presented to the
user. Demonstration:

[zacheiss@never-never-land] ~$ /bin/sh
$ (sleep 120 ; kill -ALRM $$ ) &
10992
$ kill -9 $!
$ 
10992 Killed
$ 

   So, this is normal output from the patch being applied, and should be
ignored.

   Jonathon's other problem is more interesting, although I suspect it
just means he's special, since I haven't seen the problem anywhere else.
`athinfo tow uname` indicates it's booted the 108528-13 kernel, so the
patch (or enough of it to make your machine boot) is installed, so I
suspect the patchadd output didn't lie about succeeding.

> ERROR: attribute verification of </usr/lib/mdb/kvm/sparcv9/genunix.so> failed
>     pathname does not exist
> ERROR: attribute verification of </usr/platform/sun4u/lib/mdb/kvm/sparcv9/unix.so failed pathname does not exist

   Do these files exist?  They should exist and be symlinks to the os
volume.  If they didn't, I'd suspect some prior wackiness on your
machine.  

Garry




home help back first fref pref prev next nref lref last post