Hey folks, Here's my latest Ganglia problem...this is Ganglia 2.5.4. I had a small cluster of 10 machines working just fine with this gmond.conf (only showing the uncommented lines):
name "RedHat 6.2 build cluster" mcast_channel 239.2.11.72 trusted_hosts 172.16.229.73 num_nodes 255 ...over the weekend, one of these 10 machines died with a bad hard drive. I reimaged the machine this morning form an identical disk image and a brand new identical hard drive. Now this machine only sees itself in 'gstat --all'. The other gmonds in the cluster see it, but it can't see the rest. Looks like it's broadcasting, but not listening. I've restarted gmond on all the nodes. Anyone have any ideas? Nothing would have changed on the network switch unless it somehow got reset without any human intervention. Any ideas are much appreciated. BTW...I'd like to upgrade to Ganglia 2.5.5 in the next couple of weeks if I can find the time...are there any issues with saving rrd files going from 2.5.4 to 2.5.5? Looks like it should all work fine, but thought it prudent to ask. I'm going to do some experiments with trying to use Ganglia with unicast only as well. I'd like to get away from multicast if possible. Steve Gilbert Unix Systems Administrator [EMAIL PROTECTED]
