>> On Wed, 19 Mar 2008 10:39:11 +0200, Marco Bezuidenhout <Marco.Bezuidenhout 
>> AT T-SYSTEMS.CO DOT ZA> said:

> There will be no tape devices available for the first DR test. So a
> temporary disk subsystem has been allocated to us for restores at the DR
> site. So we need to get all data on this disk and we have to be sure
> that all the data are available.

Color me incredulous.

You can't get the right equipment for your DR test, so you're going to
engage in a lot of special pre-DR activity, to further distance your
DR test from actual DR conditions.  That'll be good when the meteor
that's going to hit your building sends an engraved notification a
week or so ahead.

I advise taking the week off and going to the beach instead.  It'll
help your DR preparedness more.


Some of the most important factors in your actual disaster recovery
behavior will be related to the devices from which you will recover,
and how scattered/collocated your data is.  You are already familliar
with the problems in not having the real devices.  Doing a special
backup to capture a full image in a convenient package invalidates the
other aspect of your test.

The exercise you are describing is not going to give you useful
training on anything but best-case planned recoveries.  This is not
completely devoid of value, but as a 'disaster recovery test' it is a
charade.


If you permit your management to believe that the test you describe is
representative of real DR response times, you are going to be
absolutely slaughtered when there's a problem that doesn't give you a
week to prepare.  Avoid, avoid, avoid.


- Allen S. Rout