On Tue, 2007-12-11 at 16:07 -0500, Matt Ayres wrote:
> Cliff Wells wrote:
> > I've got the following hardware:
> >
> > CPU: 2x1.6Ghz Opteron 242 (may upgrade to a pair of 2.4Ghz dual-core
> > 880's at some point)
> > RAM: 8GB Dual channel DDR-400 ECC (may upgrade to 16GB at some point)
> > Swap: 2x16GB partitions, one enabled
> > Disks: 6x160GB 7200rpm SATA-160 on 6-channel LSI MegaRaid (64MB cache)
> > as 800GB RAID-5
> > ...
> >
> > I'd like to hear any recommendations people might have for limits to put
> > on this type of server.  I don't expect graphs or science, just the
> > voice(s) of experience =) 
> >
> >
> >   
> 40 VPS's max is what I'd put on it.  With decent load in each you could 
> find yourself with as low as 20.

I think what I've settled on is to use "vzsplit -s 0 -n 100", but limit
the actual number of VE's to 50.  This should keep the number of actual
processes reasonable and simultaneously allow me to have sufficient room
to expand the resources for each VE if needed.

And of course I can always go to the dual core CPUs, add another RAID
card and more spindles and more RAM if needed (the additional RAM should
at least help the host with caching and avoid swapping if nothing else).

Thanks for the advice.

Regards,
Cliff