LIFE IN HELL:My thirteen days in Exchange Hell

[Stewart Stremler]

begin  quoting Andrew P. Lentvorski, Jr. as of Tue, Oct 12, 2004 at 01:24:30PM -0700:
> 
> On Oct 12, 2004, at 8:20 AM, David Moreno wrote:
> 
> >  I actually read the article.  In between the lines -  the server was 
> >running for several years (YES, SEVERAL YEARS NON-STOP) in a 
> >non-technical shop who did their own backups.  They never tested their 
> >recovery scenario.
> >
> >Rule 1 - Your recovery is only as good as your planning and testing.
> 
> That is true.  But the tools don't *have* to make it hard.
> 
> In addition, the fact that the backup/migration procedure is so 
> horribly painful means that these folks left the Exchange server to rot 
> because they are deathly afraid to change anything on it.  That's not a 
> good thing either.

That might not be due to Exchange.  At, er, job[-3], they had a
sparcstation set up in the corner running their email via uucp. Nobody
messed with that machine, because it was an otherwise all-M$ shop. This
UNIX stuff was scary, spooky, and slated for replacement. Besides, it 
Just Worked, day in, day out...
 
(Naturally, it was replaced by a couple of NT machines. They were New!
Shiny! Fast! machines... *I* didn't mind, as that meant I got to put 
the Sun box on _my_ desk...)

> Personally, though, I don't have much sympathy for the individual.  
> RAID 1 is so pathetically cheap that a disk drive failure should 
> *never* cause downtime anymore.

Presumably, then, we should start seeing it standard on mass-market
motherboards.

I'm led to believe that high-end Sun machines can hot-swap *anything*,
from disk drives to power supplies and CPUs. Perhaps if downtime is an
issue, businesses should start investing in the sorts of machines that
let them avoid downtime...

...and then it might be wise, even so, to *schedule* downtime so as to
TEST the procedures and policies regarding what to do when things Go
Wrong.

-Stewart "Something will always go wrong. Eventually. Somewhere." Stremler