scyld slave node problems

[Donald Becker]

On Thu, 11 Apr 2002, Stepan Kruglikov wrote:

> I solved this problem by increasing memory on each node to 64MB. You can
> also get the node up and running with 64MB, setup node partitions, and then
> delete ram drive and run with 32mB. Although it works, I recommend doing it
> only in case if you are interested in proof of concept cluster.

It's possible to trim the cached library list in /etc/beowulf/config and
fit into 32MB.  But only the most trivial application will run with 32MB
and no local disk.

-- 
Donald Becker				becker at scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993