1st of all, what happens if your current (SBS) server crashes while
you read this comment? Do you have backups and a means of restoring
it? If yes, skip the next 3 paragraphs, but if no, read on. This is
the first thing I will recommend you do! If you don?t have the
expertise/equipment to do this, best you hire someone (with equipment)
to do it for you. Keep two copies, one of which should be kept
off-site!
In fact even better, do a risk assessment first. If you are unfamiliar
with this procedure, in simple terms, you ask yourself some really
nasty questions and try answer them. Determine the likelihood of each
event and factor that into the equivalent fiscal measure of such an
event. For example: Global Nuclear War: likelihood is 1 in 1000 years
(my guess!). Financial implications: Total irrecoverable loss of $15
million. Multiply (1/1000) x 15,000,000 = 1,500. However ridiculous
this example, it clearly parameterises the ?risk factor?.
Let?s take a much more relevant example. Single primary server goes up
in smoke. There are no backups. Estimated business downtime = 7-days
with a further 30-days worth of slow go, while data is being captured
manually from paper records. Let?s estimate this at an arbitrary cost
of $40,000. Let?s say the likelihood is 1 in 5 years. The result is
8,000. This is clearly a higher risk than Global Nuclear War! Let?s
leave it at that, I am sure you already know this or at least are
catching my drift.
In your risk assessment exercise, be sure to include the possibility
of hiring an expert that turns out not to be an expert. This scenario
can easily match up to the above example, with or without the smoke
effects! I say this as I suspect you do not have a suitable
professional in mind! This is really good for you. How many years have
you been operating without having to call in a consultant? Sounds like
you have a winning arrangement going! There are ways to reduce your
risk of exposure however. One of which is to use your in-house people
with proper safety and controls to protect you from disaster. The
point here again, whoever does the job, you want those backups first!
All said and done, I would use my in-house techs if I were you. You
don?t have immediate time pressures. It will be a learning experience
to them and you will have much better in-house support in future. All
can be done without even touching your SBS server at all! I stand to
be corrected on this, but I don?t think SBS is all that compatible
with ?running in the background?. On the other hand, W2K does have
these features built-in (IntelliMirror) and I recommend you rather get
a 2nd W2K server and do a switchover when your W2K network is fully
functional. You can then keep the SBS on the shelf for a while if for
nothing other than emotional piece of mind! Always a good thing to
have, but again, keep it off-site, but secure!
If you decide to take the in-house route, all the technical info is in
the W2K help files. It is in fact a comprehensive manual and takes a
little while to get used to, but it?s all in there! Get your techs to
install W2K on the new server and let them play with it for a while.
After a week they should be able to introduce client machines and
eventually they should be able to access the SBS network and suck all
the files from that server. You should then have your W2K network
close to where you want it to be. Get them to backup everything and
then restore the lot onto the 2nd server. Remove the first server and
run the network on the 2nd server. Re-introduce the 1st server and
then start introducing peripherals. You can do the printers one by one
after hours, until all driver issues are resolved.
Before you know it, you will have a fully functional W2K network with
IntelliMirror and all! Your risk factor will have gone down
significantly and you will feel a lot more confident. I have upgraded
from SBS (25 users) to W2K about 2-years ago and have only praise for
W2K. Until about 1-week ago I never ever had the W2K server crash on
me. It was a memory fault that caused it, one of the 256Mb modules
died. I found the fault by trial and error. None of the diagnostic
programs could detect the memory fault.
These are my comments, other views are welcome and will be respected! |