I am currently looking at using Ganglia 3 to monitor compute clusters
running Solaris 9, Suse Linux 9, and AIX 5.2 ml 3 on the Power4 and 5
architecture. Everything compiled nicely on the Solaris and Linux servers.
After a little work, gmond and gmetric compiled and start correctly on the
AIX servers.
There is one strange thing- I am seeing only partial metrics being
reported for the AIX servers, and some of the ones that are reported (such
as cpu_user) seem to be an inverse. Has anybody else seen this behavior
before? If so, is there a patch?
