only zero values in rrd - other hosts work great

[Thalamus]

Hi everyone !

I have a problem that I was hoping you could help me solve. I have several redhat servers that I poll using a template that I've found from this forum. All servers except one works flawlessly. There is one host however - that I poll and get cpu, load and the rest from - but the interface polls are returned as 0 values. If a do a 'snmpwalk <server>' for both 32 and 64 ... I get these responses.

IF-MIB::ifHCInOctets.1 = Counter64: 2044201761780
IF-MIB::ifHCInOctets.2 = Counter64: 296299310823

IF-MIB::ifInOctets.1 = Counter32: 4092296180
IF-MIB::ifInOctets.2 = Counter32: 4241534695

All these servers run the same version of net-snmp, are on the same patch level. RedHat 5.6.
I've deleted that host, recreated it several times. Dumping rrdfile shows me :

<!-- 2011-03-23 10:40:00 CET / 1300873200 --> <row><v>0.0000000000e+00</v><v>0.0000000000e+00</v></row>
<!-- 2011-03-23 10:45:00 CET / 1300873500 --> <row><v>0.0000000000e+00</v><v>0.0000000000e+00</v></row>
<!-- 2011-03-23 10:50:00 CET / 1300873800 --> <row><v>0.0000000000e+00</v><v>0.0000000000e+00</v></row>

I could not find any solution to my problem - not even in Gandalf's debug guide did me any good.
Any Cacti guru that might have a clue to why this is happening ?

[niobe]

Are these currently counters currently higher than on the other servers? Unlikely to be an issues with size though.
Nice to include the RRD dump, but go back one step. If you set cacti to DEBUG, what does the poller line return? 0 would be odd, U is more common when the poller can't get a result it considers usable.

[Thalamus]

Are these currently counters currently higher than on the other servers? Unlikely to be an issues with size though.

No, they are as expected. I noticed now however that on my troublesome server, It seems to be the localhost interface that actually are increasing. Come to think of it - not long time ago I noticed that this host did not have a ref for eth0 mapped towards its hostname. This server has not been restarted since then. I now just wonder if this might be the real reason. SNMP daemon has been restarted but not the services themselves.

I will investigate and come back with my findings.
Thank you so much for the input - you made me think in a whole different direction

[Thalamus]

Graphs get generated like expected on loopback interface. So, my conclusion is that I'll have to recheck this host once we have our next scheduled downtime.
Thanks again for any help - appreciated !