On Mon, Aug 06, 2007 at 09:21:13AM +0200, Thomas Bruesemeister wrote:
> our WebMDS complains about missing aggregator
> data for the Multi Job GRAM in GT4.0.5.

What's the error?

> Anyway publishing Ganglia information for every MJFS
> produces *a lot* of overhead for big clusters. What
> do you recommend?

I recommend configuring Ganglia how I mentioned previously in this
thread.  Only scheduler information should be published for each MJFS.
How did you configure it to see repeating Ganglia (cluster) data?

-Neill.

Reply via email to