RE: Opinons?

> From: Tom Jedrzejewicz
> 
> You will be much happier if you get two separate boxes, one for the
server
> functions and one for workstation/WDSc functions. Buy a good ThinkPad
(<
> $2k) or a good ThinkCentre machine (<$1500) with a bunch of RAM for
the
> WDSC work.

We'll see.  I already have a nice machine for development -- 3GHz P4,
1GB RAM, 7200RPM RAID 1 drives.  My guess, though, is that the 15K
drives will be a huge boost to WDSC.

The other issue is that the server functions are for development (i.e.
testing) not production.  Thus, they won't be in use at all times.  I'll
do some development in WDSC then deploy it and test on the servers.  I
don't foresee the servers and WDSC competing.  I'm pretty comfortable
that this use will work.  If not, I go back and beef up the old
workstation and dedicate the xSeries to serving.


> Be aware that not all apps can use dual processors .. I don't know
about
> WDSc .. and the Windows installation is different for a dual processor
> system. Also, the things that make a system a good server aren't the
same
> things that make it a good workstation.

"The Windows installation is different for a dual processor system"?
What do you mean by that?

As to "the things that make a system a good server," I've thought about
that, and I'm actually a little confused.  In the old days you often had
to choose between fast disk and fast CPU, but with a configuration like
this, I'm getting everything.  What exactly do you consider things that
make a good server that wouldn't apply to WDSC?  It seems that really
fast disk drives, lots of memory and powerful processors apply to either
one.  What am I missing?


> Also, do you really want to be developing on the same box hosting even
the
> low volume web sites? What happens if you crash it, or have to
re-install
> WDSc, or one of the other nightmares that developers periodically get
in
> with their systems .. are you prepared for these web sites to be down?

Yes.  These are meant to be experimental, development or testing sites.
If it turns out that one of the sites needs more stability, I'll get
another server.


> My thought has always been that once I suck it up and decide to buy, I
buy
> the best CPU you can afford, generally 1 step below the leading edge.
> Memory
> and disk can be added later for incremental costs, but replacing a
> processor
> is a nightmare. What you are effectively doing is buying a few months
of
> longer life on the back end.

The 3.2GHz is perhaps two steps below, but the cost seems right.  It's a
flip of the coin between it and the 3.4GHz machine.  The 6% processor
increase doesn't seem justified.


> As an aside .. I see nothing related to backup on this configuration
...

My current backup policy is external USB drives.  They work like a
charm.  But I'm interested in other options.  However, I won't buy a
$17,499 tape drive to back up a $3000 server <grin>.