> > Yes. MRTG has no way to uniquely identify which interface you want > to graph. > > The one (and only one) that's "up" of course. :-) > > > Many are called 'ppp0', many have type 23, the ppp interfaces have no > MAC address, and so on. > > Indeed. But only one is up.
My view is that the problem lies with SNMP (and not with MRTG) since there definitely should not be multiple instances of ppp0 in SNMP. In this case, changing your MRTG version will not help. MRTG cannot simply say 'grab the only one thats up' since in many cases an monitored interface may legitimately be down and should still be checked. You could, of course, write a data collection plugin to have this functionality, though. I'm not sure why SNMP is misbehaving, particularly if (as you say) there has been no software update or changes since the time it last worked. It might be worth trying to identify what the event is that creates a new SNMP interface table entry as this is simply pulled from the kernel and indicates a resource problem there. It maybe the problem is being caused by the other end of your PPP link, or by options on the link for reconnection? This is getting a bit beyond me, though. Steve _______________________________________________ mrtg mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/mrtg
