[collectd] help with collecting # of accepted BGP routes

matthew sporleder msporleder at gmail.com
Sun Oct 18 04:35:53 CEST 2009


On Wed, Oct 14, 2009 at 2:24 PM, L. Gabriel Somlo <gsomlo at gmail.com> wrote:
> Hi,
>
> I'm trying to graph the number of accepted bgp prefixes for
> each peer of my border router. Using snmpwalk, I can retrieve
> counts for all applicable BGP neighbors and protocols
> (e.g. unicast, multicast) by running:
>
>
> # snmpwalk -v2c -c comstr router CISCO-BGP4-MIB::cbgpPeerAcceptedPrefixes
> ...
> CISCO-BGP4-MIB::cbgpPeerAcceptedPrefixes.192.168.16.106.ipv4.unicast = Counter32: 190837
> CISCO-BGP4-MIB::cbgpPeerAcceptedPrefixes.192.168.16.106.ipv4.multicast = Counter32: 4477
> ...
>
>
> or
>
>
> # snmpwalk -On -v2c -c comstr router CISCO-BGP4-MIB::cbgpPeerAcceptedPrefixes
> ...
> .1.3.6.1.4.1.9.9.187.1.2.4.1.1.192.168.16.106.1.1 = Counter32: 190835
> .1.3.6.1.4.1.9.9.187.1.2.4.1.1.192.168.16.106.1.2 = Counter32: 4477
> ...
>
>
> Using collectd, I added the following to the snmp plugin's config file:
>
> <Plugin snmp>
> ...
>   <Data "accepted_bgp_pfx">
>      Type "current"
>      Table true
>      Values "CISCO-BGP4-MIB::cbgpPeerAcceptedPrefixes"
>   </Data>
> ...
>   <Host "router">
>      Address "10.10.1.136"
>      Version 2
>      Community "comstr"
>      Collect "traffic" "errors" "unicast" "net_mem" "net_cpu" "net_temp" "accepted_bgp_pfx"
>      Interval 300
>   </Host>
> ...
>
>
> I get .rrd files for everything else under ...router/snmp/*, but nothing for
> "accepted_bgp_pfx". Also, no errors logged by collectd.
>
> I'd expect to see something like current_192_168_16_106_1_1.rrd,
> current_192_168_16_106_1_2.rrd, etc. for each peer/protocol.
>
> Any ideas about what I might be doing wrong ?
>

You may need an instance or instanceprefix to convince it to iterate
and store everything.



More information about the collectd mailing list