Hardware plays a big role in OS reliability and security. A few years back there was a report that IBM had a mainframe running continously for 7 years. When VM has the ability to do VM failover then the uptime will rise dramatically on x86 platforms.
VMWare's ESX Server, which I have not yet had the good fortune to be able to play with, is getting closer and closer to this on current x86 hardware, but really shines when you're running the "system" on a blade or cluster node backed by shared storage. Currently, you can do live snapshots of running system images, migrate the snapshots to another node, kill the original and fire up the snapshot, and you're back in business. I think they actually preserve the full machine state at the snapshot, which means you can have very small downtime failovers. I was going to try to get a demo of this capability over at Warner Bros. IT last time I was up in Burbank, but they had a company holiday and the timing didn't work out.
Granted, if the node you're running system images on happens to get toasted, you're still SOL until you migrate the image to another node, but it's better than having to do a restore or a complete re-install.
Still not as good as IBM's VM stuff, or any vendor's who's VM tech is integrated into the hardware, but it will hopefully lead in that direction.
Gregory
-- Gregory K. Ruiz-Ade <[EMAIL PROTECTED]> OpenPGP Key ID: EAF4844B keyserver: pgpkeys.mit.edu
PGP.sig
Description: This is a digitally signed message part
--
KPLUG-List mailing list [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
