Hi, I'd like to know what's the best topology for a unicast deployment. I want to monitor several clusters (memcache / redis / mysql / etc...).
--- One solution is to configure each node in a given cluster to talk to a given head gmond (or two for HA) M1, M2, M3 -> send to MHEAD:MPORT R1, R2, R3 -> send to RHEAD:RPORT ... a gmetad is configured to get data from all the head gmond MHEAD:MPORT, RHEAD:RPORT, etc datasource 'memcache' MHEAD:MPORT datasource 'redis' RHEAD:RPORT since all my head gmond and my gmetad are on a single box, this is a lot of gmond processes listening to different ports (handled by supervisor) --- The other solution is to configure each node in a given cluster to talk to each other: M1 -> sends to M2,M3 M2 -> sends to M1,M3 M3 -> sends to M1,M2 R1 -> sends to R2,R3 R2 -> sends to R1,R3 R3 -> sends to R1,R2 ... And have gmetad talk to a given machine for each cluster (adding all the others for HA) datasource 'memcache' M1, M2, M3 datasource 'redis R1, R2, R3 This is more chatty. Are both solutions acceptable or is there some sort of consensus on what to use? Also I have data send by gmetric and some time ago it seemed that the data where send to the send_channel but not localhost. This could be a problem in the second scenario where we don't get the data for the first host. Is that an expected behavior of gmetric, a bug or did I do something wrong? Thanks --Gilles ------------------------------------------------------------------------------ Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 _______________________________________________ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
