I've got a cluster of hosts running gmond with identical configs using multicast. I'd like them all to be one cluster, but instead, two hosts, hostfoo and hostbar, are acting as a separate cluster. If you query either of their gmond's, you get metrics for hostfoo & hostbar only; if you query any other gmond in the cluster, you get metrics for all hosts *except* hostfoo & hostbar.
I believe I have narrowed this down: Every host in this cluster, in addition to its primary interface, has an additional VLAN-tagged virtual interface (on the same physical network & switch). So for example eth0, and eth0.2. It seems that gmond in some hosts is sending its multicasts via the VLAN tagged virtual interface, while on other hosts it's using the untagged interface. So even though the multicast packets are all on the same switch, they're acting as two disjoint networks. I've tried controlling this by putting mcast_if = eth0.2 and/or mcast_if=eth0 in my udp_send_channel sections, but it doesn't seem to make a difference. How can I control which interface gmond uses? Update: I have found a different workaround: Although I can't seem to control which interface gmond uses for the send channels, I *can* use mcast_if in the udp_recv_channel. If I set up two receive channels on each host, one with mcast_if=eth0 and the other mcast_if=eth0.2, then they get metrics from all of the other gmonds. Still interested in the answer to my original question, if anyone knows. -- Cos ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
