Hello there,
I've read the previous thread about scheduler O/(1) patch.
As i'm in a study process for vserver,
i've found a good document that speak about per Vserver fairness in the
E.Kotsovinos Report:
http://www.cl.cam.ac.uk/~ek247/research/fyr.pdf
As we talk about scheduler algorithms,
my first question is: Can we achieve a good per Vserver Fairness with a
O/(1) patch ?
Second question: What about a great feature for vserver ctx patch if we
can control some cpu limits per Vserver ?
As this also scheduler related ...
hugh
