SMP robustness

[Greg Lindahl]

> i've been thinking about HA solutions and realized that don't know what
> happens if one of SMP machines' CPU fails. Is it different in Intel, IBM
> or SUN SMPs? Does machine hang ?

If one fails, the entire machine fails. The point of clustering in HA is to
have an entire machine to take over the work of the failed machine.

As you can imagine, big SMPs like an SGI O2000 are pretty fragile in this
respect.

-- g