Hello All I am running Ganglia 3.0.1-1 on a 90 Node Linux Cluster Beowulf style.
I have an issue that is driving me nuts. I have the same gmond.conf file running on all 90 compute nodes. But, when I run the gmetd front end it appears to see only half of my cluster..ie 45 nodes. The way I hoped to get around this was listing several ip addresses for the gmonds that could be polled for information. So I put in the IP address of every 10ntth compute node as my "data_source" When I do this the gmet frontend still only "sees" 1/2 of the compute nodes. Now and then it will "see" all of the compute nodes, but this is quite rare. How do I fix this apparent problematic gangila behavior? Thanks! Dan Roberts
