Netscaler interface graphs have identical data

[net-guy]

So I've searched and troubleshot this, and I still can't figure it out. The logs of cacti indicate spine is polling the correct OIDs, but it comes back with the same data every time. I have two 10G interfaces that are *not* identical in throughput over a 5 minute average.

The logs indicate two different OIDs are being polled:

02/27/2010 02:05:13 AM - SPINE: Poller[0] Host[469] DS[44202] SNMP: v1: <name removed>, dsname: ns_traffic_out, oid: .1.3.6.1.4.1.5951.4.1.1.54.1.7.3.49.47.49, value: 835093221447
02/27/2010 02:05:13 AM - SPINE: Poller[0] Host[469] DS[44203] SNMP: v1: <name removed>, dsname: ns_traffic_out, oid: .1.3.6.1.4.1.5951.4.1.1.54.1.7.3.49.47.50, value: 835093221447

Yet the same data comes back. If I manually SNMP walk the OID, this is not the case.

<i can individually hit both as well, here is the index snippet>
[root@]# snmpwalk -v2c -c <removed> <removed> .1.3.6.1.4.1.5951.4.1.1.54.1.7.3
<sniP>
SNMPv2-SMI::enterprises.5951.4.1.1.54.1.7.3.49.47.49 = Counter64: 1067370442847
SNMPv2-SMI::enterprises.5951.4.1.1.54.1.7.3.49.47.50 = Counter64: 1037249544042

The counters are always different, and never match -- these oid's equate to both 10G interfaces on the netscaler.

I think it has something to do with the snmp_query regexp, and i've adjusted it -- but can't seem to figure this out for the life of me:

The relevant portion follows. The REGEXP started with 31, but I got n/a for everything until I moved it to 30. Now I get duplicate numbers for everything.

Obviously I'm having some sot of a counting problem and can't do basic math or my understanding of the template is flawed. Is this my problem?

<name>Get SNMP NetScaler Interfaces</name>
<description>Queries a NetScaler 9.0 for a list of monitorable interfaces</description>
<oid_index>.1.3.6.1.4.1.5951.4.1.1.54.1.1</oid_index>
<oid_index_parse>OID/REGEXP:^.{30}(.*)</oid_index_parse>

<fields>
<ifName>
<name>Interface Name</name>
<method>walk</method>
<source>value</source>
<direction>input</direction>
<oid>.1.3.6.1.4.1.5951.4.1.1.54.1.1</oid>
</ifName>
<ifRxBytes>
<name>Number of bytes received by this interface</name>
<method>walk</method>
<source>value</source>
<direction>output</direction>
<oid>.1.3.6.1.4.1.5951.4.1.1.54.1.3</oid>
</ifRxBytes>
<ifTxBytes>
<name>Number of bytes transmitted by this interface</name>
<method>walk</method>
<source>value</source>
<direction>output</direction>
<oid>.1.3.6.1.4.1.5951.4.1.1.54.1.7</oid>
</ifTxBytes>



So I've looked at the poller cache, it has the right complete OID. The interface names are right - the log even indicates it's polling the full OID -
yet it comes back with the same result for all interfaces. What am I missing here? Where can I start next?

[Linegod]

You are using SNMP v1 in Cacti and SNMP v2c at the command line.

Try fixing that...

[net-guy]

Thanks - but I've already tried both ways (that was an initial troubleshooting step). Both ways give the same result, including change versions in cacti and on the CLI. CLI snmpwalk always results in correct information regardless of snmp version. Cacti always shows the same.

quote="Linegod"]You are using SNMP v1 in Cacti and SNMP v2c at the command line.

Try fixing that...[/quote]

[Linegod]

Maximum value in the template set/set to low?

[net-guy]

That was a good idea; I didn't even bother checking that. I just did, it's set up to 10000000000 -- so plenty of room. Good idea though. I'll probably have to sniff/trace exactly what cacti is doing unless anyone else has any other ideas?

Maximum value in the template set/set to low?