--- Alexei Rodriguez <[EMAIL PROTECTED]> wrote: > Hello and thank you for the reply. Responses inline: > > > > - which version of ganglia/gmond are you running? If possible, > please > > try out 3.0.2. > > I am running 3.0.2. >
too bad :-) > > - are you using the first or the second of the two NICs? > > I am using the second NIC, but they are different chipsets. This is > on > the Sun X2100 server, which has 2 built-in NICs: nge0 (primary, > nvidia > based) and bge0 (scondary, broadcom). If I use the primary (nge0), > then gmond reports stats for it. However, we are using bge0 as the > network interface on our systems. > OK, this is interesting. Maybe the bmc driver does not interface well with "kstat". See below. > > - how are your NICs named? The code drops everything starting with > 'l' > > or 'o'. > > nge0 and bge0. > that should be fine. Actually, I was wrong about the filter. The code just throws away any "lo"cal interfaces. > > Unfortunatelly Solaris/AMD64 might not be very well tested? > > I suspected that could be the case. Hopefully > more of us will be using it and that can change :) > I have the impression that you may not be wrong here :-) AMD64 looks like a very interesting platform for replacing Sparc. So we have to fix our bugs. > > You can try to put some debug statements into the > "extract_if_data" > > routine. > > I'll give that a spin. > One thing to check is the output of the "kstat" command. We are using the kstat(3KSTAT) API and if that does not give the right numbers, gmond is left in the dark. So, please do: kstat -n bge0 and look for "ipackets64". Does it show changeing numbers? Cheers Martin ------------------------------------------------------ Martin Knoblauch email: k n o b i AT knobisoft DOT de www: http://www.knobisoft.de
