On Fri, 6 Feb 2015 05:35:43 PM Prentice Bisbal wrote:

> Do any of you disable swap on your compute nodes?

We run diskless clusters these days and so we don't get a choice. :-)

Seems to be fine - we use Slurm's cgroups support to constrain memory usage and 
don't overcommit memory.

Java's been the only issue (defaulting to trying to allocate 1/4 the RAM of 
the node on startup) but Java 8 finally fixes that problem and so it'll now 
start in our default allocation of 2GB per core without any extra command line 
options.

All the best,
Chris
-- 
 Christopher Samuel        Senior Systems Administrator
 VLSCI - Victorian Life Sciences Computation Initiative
 Email: [email protected] Phone: +61 (0)3 903 55545
 http://www.vlsci.org.au/      http://twitter.com/vlsci

_______________________________________________
Beowulf mailing list, [email protected] sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Reply via email to