[CUBE] Strange error message

Phil Mitchell pmitchell at databanque.net
Thu Jan 29 01:48:09 PST 2004


On Jan 29, 2004, at 12:00 AM, J.C. Webber III wrote:

> I don't mean to be harsh either, but this is the typical support
> engineer's response.  Get it fixed and the customer off the line
> as quickly as possible.
>
> Never mind that this 'shotgun' approach removes any possiblity
> of discovering the root cause of the problem, thus providing
> the user with some clue as to how to avoid a reoccurance.
>
> A complete scratch install should be the last resort, when you've
> effectively given up, not the first thing you try.
>
> -- jcw

Well, J.C., the objective of a consultant IS just that - to get the 
client/user back into production ASAP! Perhaps Laurie A Duncan of NYC 
can simply and easily diagnose the cause of a kernel panic remotely, 
but I cannot. Further, I have never isolated THE SOLE cause of a KP! 
There seems to be a myriad of possible causative actions and issues. I 
suggested an archive install in this instance because I was not there 
to observe, and:

A. The install only takes 20-30 minutes for the average user on a 
G-series machine with adequate RAM and hard drive space, then an hour 
or so (with broadband access) letting software updater do its thing and 
you're good to go again...

B. Users, preferences, network settings, etc. are brought over and your 
old systems are safely retained in the "previous systems" folder - 
should you need a pref, a font, etc...

C. I have NEVER had a KP situation not immediately cured by an archive 
install...

I can also say that the majority of our clients/users (including 
everyone in all our offices!) of OSX for a year or longer has had one 
of these anomalies! I personally have had two on my Cube and wasted 
many, many diagnostic hours on the first one - with no conclusive 
results. The second time, I just took my own advice and was back up and 
running again in a couple hours. I still have no clue as to what set of 
circumstances perpetrated either KP, as both occurred while I was 
actively engaged in production work - no new software, hardware... 
nada!

In closing, let me say that when a production user is unable to work 
because of a "down" computer system, their mission directive to the 
consultant is: FIX IT! DO IT NOW! Would they like to know the cause - 
if only to avoid re-creating the problem? Certainly! Do they want to 
know badly enough to justify hours spent analyzing logs, running 
diagnostics and explaining all their actions leading up to the 
crash/problem in detail? They do not.

Apologies to all for the length,

pm



More information about the Cube mailing list