[ILUG] Have you had your daily dose of reality?

[Paul Jakma]

On Fri, 18 Feb 2000, John P. Looney wrote:

>  True, but you can cluster in the box. You can say "Those 8 motherboards
> are one 32 CPU system, and those 8 motherboards are the other half of the
> cluster". You then get Oracle PDB (parallel database), and if one part of
> the E10K gets a fault, it'll reboot the OS on that virtual machine. Oracle
> will continue along slowly for five/ten minutes, at half-speed, till the
> rest of it comes back up.
> 

a lot of tru64 cluster is in app space aswell. i heard that most of the
gkn tru64 cluster course is about the scripts that control it. :)

however. once, the app has failed over, and the faulty system board
isolated: can you change that system board and bring it back up without
rebooting the other domains?

>  It's only application level clustering, but still, not bad.
> 

larry mcvoy had a post on l-k a while back. according to him single
image parallel hardware systems (like the e10k> are bad. he reckons a
large cluster of single image systems that can share resources are the
way to go, rather than finely locked complicated smp kernels.

one idea he put forth was an "SMPfs" a sort of nfs for kernels in a
cluster to claim locks on resources in other machines. Sounds rather
cool.. eg kernel on host1 could claim memory on host2 via smpfs. that
kind of thing. So benefits of single system for apps's, without most of
the complexity of a single system kernel.

he has papers on www.bitmover.com about it. should be a good read.

> Kate
> 
> 
> 

regards,

Paul Jakma.