diskless clients? beowulf-newbie seeks advice

[Serguei Patchkovskii]

On Fri, 22 Jun 2001, Brian LaMere wrote:
> well, we do use 15k rpm scsi drives, but yeah....ya want to do as few reads
> as possible even then.  Just seems with a potential of there being 100 of
> these nodes, I very well may have to boot one every few days...and that
> would be lots of traffic.  

Not all that much, actually. We run a cluster with 93 diskless (well, actually
OS-less; they do use local disks for scratch etc) nodes. Booting all the nodes 
simultaneously (usually after a power failure) pretty much saturates two 100Mbit
links to the boot server for about 15 minutes. This comes down to about 20Gbytes
of data being transferred altogether - nothing to really fret about. Incidentally, 
this is with Tru64 - which is substantially more bloated than Linux. During 
normal operation, the traffic to the OS server is much, much lighter ...

/Serge.P

---
Home page: http://www.cobalt.chem.ucalgary.ca/ps/