
Originally Posted by
esilverst
Our clients have installed a 3.2 DE cluster to stress test their application. They are running heavy duty experiments on this cluster of 13 machines.
They found that 2 out of the 13 machines that participate in the cluster had processes (Lotus Notes and Explorer) killed while a heavy job was running on their cluster.
Has anyone seen something like this happening? What could be the reason?
After we asked them to bump down to 1 cpu for each of the agents, we have not seen any more incidents like this. The original configuration had each machine contributing one half of its total cpus to the cluster. But now even the 4-core machines are only contributing 1 cpu.
Does Symphony 4.0 DE support host scavenging? Or is that only available with EGO?