ctacke/>
Tue Dec 18 06:21:14 PST 2007
The reason we have so many disks (so a diff disk based of a diff disk based
on a diff disk, etc) is that it allows us fine granularity for creating new
images and what their contents will be. If we need to test something
against PB with no QFEs, it's easy. If we find suspect that a refresh of
Studio caused issues we can test it readily and easily. It also allows us
to build a "fresh" machine with hopefull the exact config we want very
quickly.
When we want to apply new QFEs, we create a new diff disk against the
previous QFE VHD and then a new "usable" machine that uses a diff disk based
off of that.
Yes, it seems complex, but it's the fastest way we've found to bring up new
machines for any given environment. I've not seen any real expense other
than disk space, and I'm not sure a diff disk is any worse than a standard
one. VPC files are inherently huge.
--
Chris Tacke, eMVP
Join the Embedded Developer Community
http://community.opennetcf.com
"Remi de Gravelaine" <gravelaine at aton dash sys dot fr> wrote in message
news:uHbfZRVQIHA.4684@TK2MSFTNGP06.phx.gbl...
> Hi Chris,
> Thanks for your answer.
>
> Your scheme looks clever, albeit probably a little bit overkill for what I
> have to do, id est almost always working at the PB level.
> I am not sure I understand correctly the concept of "hierarchical VPC".
> Does it just mean that you install Virtual PC in virtual PCs? If so, is
> there any drawback or performance penalty?
> Do you spawn a new VHD each time there's a QFE? I can't figure out if it
> is an expensive operation or no. What is the cost of a differencing disk?
>
> Sorry for so much questions and thanks again.
> Remi
>