> > The difference between Sprite and Linux is that linux WORKS. You could
> > also say that Solaris already does what Linux DOES. That is not the point
> > ... can you imagine using the linux kernel on a number of PC's to build a
> > mainframe style expandable cluster.
>
> There are some funny special cases that need to be considered. How
> does your proposal handling failing components in the cluster? Imagine
> a network failure that split your cluster in two parts, each fully
> functional but unconnected to the rest. The so called split-brain
> syndrome. Now each cluster half will continue processing and assumes
> it is authoritive? Imagine a database system being split up into
> two systems ...
the problem is that there is no 'absolute' authority function available.
So we should simply ignore this problem.
think about it, you CAN have physically valid special cases when it's
totally valid to have two separate systems. (please dont start another
light speed war, but if you put such a cluster slightly below and above
the event horizont of a black hole you will have two totally separated
physically systems).
so the solution is to define 'trusted' authoritive media (an additional
network of serial lines or something similar), and swear loudly when a
split happens despite these measures ;)
not even human brain can resolve such conflicts ;) [=schizophrenie]
-- mingo