Hello Ganglia users and devels, I have a deployment of Ganglia 3.2 with two monitoring servers and ~300 clients. Each server runs gWeb2, a single gmetad (C, not Python) and multiple gmond, on separate ports, for separating different "clusters" Mostly because of network constraints, I am using unicast, so each client points to a specific port on both servers.
The setup is working nicely, especially after I started using rrdcached, but I have this nagging issue that each time that I reboot or restart the receiver gmond, I need to also restart all the senders on the clients. Because of this I miss data in the RRDs, because gmetad finds the running (but empty) gmond, and does not failover to the other server. I have tried to set send_metadata_interval to non-zero, I do get back the metrics, but some report plots are not correct, in particular missing the CPUs in the load_report and the Total and Use in mem_report I suspect that the constant metrics (like total RAM, nr of CPU) are not being resent even with send_metadata_interval. Has someone seen this before? Is there any simple fix? It would really be nice if also the unicast receiver could request metadata like the multicast does.. The other idea that I was thinking about was to put only the receivers on the two servers in multicast - but this would mean extra NICs and a dedicated cable across the whole data room, so before embarking on such a job I'd like to hear if anybody tried it before. Cheers, Sergio ------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/ _______________________________________________ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
