Out of the blue this evening I noticed one of my data sources was
consistently showing a NaN on the graphs. I messed around for a
while with various things but could not fix it, so I deleted the
graph and all its data sources and created it again, just to see
what would happen.
What happened was, that data source began getting collected again
fine, but some other ds started failing instead.
Looking in the cacti.log I saw lots of:
Code: Select all
07/08/2006 01:20:09 AM - CMDPHP: Poller[0] Host[2] WARNING: Result from SNMP not valid. Partial Result:
07/08/2006 01:20:09 AM - CMDPHP: Poller[0] Host[2] SNMP: v1: 212.13.198.69, dsname: unicast_in, oid: .1.3.6.1.2.1.2.2.1.11.1, output: U
deleted it. Querying that OID manually with snmpwalk I see:
Code: Select all
IF-MIB::ifInUcastPkts.1 = Counter32: 519996375
cacti settings so why should this suddenly start breaking?
This is an everyday install of Debian Sarge getting interface stats
out of net-snmp. It is always that OID which I get the message
about. I suppose if I deleted the graph and datasources and created
it again then the problem would move elsewhere.
On the cacti server:
Code: Select all
ii cacti 0.8.6d-1 Frontend to rrdtool for monitoring systems a
ii php4 4.3.10-16 server-side, HTML-embedded scripting languag
ii php4-cli 4.3.10-16 command-line interpreter for the php4 script
ii php4-snmp 4.3.10-16 SNMP module for php4
Code: Select all
ii libsnmp-base 5.1.2-6.2 NET SNMP (Simple Network Management Protocol
ii libsnmp5 5.1.2-6.2 NET SNMP (Simple Network Management Protocol
ii snmp 5.1.2-6.2 NET SNMP (Simple Network Management Protocol
ii snmpd 5.1.2-6.2 NET SNMP (Simple Network Management Protocol
Andy