Thats not totally correct. The problem Leif describes pertains to Soft
RAID where windows raids only the selected partitions - with soft RAID,
the boot pointers get stuffed (boot.ini arc paths get stuffed ad
possiblyboot sectors / active partion, MBR etc. all easy to fix if you
know what to do), but can be fixed without having to fix the raid first -
or at least that is what used to happen, they could have fixed it.
With H/W controller RAID 1 the system *will* allow you to boot and carry
on (or carry on if the drive fails at run time - that is what RAID 1 is
for).
When you replace the faulty drive it *will* resynch the drives - these
should be quite automatic but each controller / firmware version is
subtally different. I have not timed the nvidia implementation of
resynching, but the Intel version runs at about 80 minutes or more per
100GB - a long time. With the Intel RAID (ICH5R / ICH6R - 865 / 875
chipsets etc.), this synch runs while windows is running resulting in
quite degraded performane for that period, but you can use the system
while this is running. Adaptec controllers are a bit better in some
respects and not as automatic in others - at least they have Alerts.
I suggest strongly that you do a trial XP install onto RAID 1, then
simulate a failure by powering off, pulling out the power & SATA (?)
connector for the second drive, rebooting then consulting the manual and
whatever happens fiugre out what to do, what impact it has, how long it
takes and so on. In between reconnecting the 'failed' disc try a boot to
confirm what I say.
It ios also important to know how you are alerted to failures: Intel
flahes a loud boot screen and shows a flasing toolbar icon when a RAID 1
fail happens, Adaptec SATA shows nothing worth mentioning at boot time,
but has email and other alerts including syslog.
I would appreciate it if you could report back as I haven't yet configured
a RAID 1 on my A8N-E and done the above - regardless its on my own to-do
list.
Have fun
HTH
