Hi, I've been playing with ganglia over the past couple of days but am hitting a wall.
I've got a small 4 node cluster I want to monitor. The web-frontend and gmetad run on host snoopy, while gmond is running on the other 4 hosts, 2 of which are multi-homed. In gmetad, I have "my cluster" set to host1, with the failover host set to host2. The gmetad host can connect to port 8649 on host1 or host2 and get output. Yet for some reason, all but host1 show as being down, when they're not. I'm running the latest versions of the various ganglia packages on all the hosts. I thought this may be a problem with multicast, but I don't know enough about it. Just in case, I set explicitly set the mcast interface, but this did not help. Any ideas? Is this too little information for anyone to be of help? Thanks, --john campbell