[collectd] snmp big values

[igor at skif.net.ua]

Mariusz Gronczewski написав(ла):
> Hi
> 2009/10/27 Янченко Игорь <igor at skif.net.ua>:
>> Hello,
>>
>> I use collectd for monitoring swiches with big traffic around 800M-1G. I
>> have next problem:
>> when I set Interval 300, collectd write into rrd complete nonsense
>> values. But when I set short interval (like 10) I get correct values.
>>
>> I also use mrtg for monitoring, for correct values in mrtg I write
>> ":::::2" after ip in config, like this:
>>
>> Target[10.10.100.5_Vl111]: #Vl111:comunity at 10.10.100.5:::::2
>>
>> mrtg have interval 300 and write correct values with this string.
>>
>>
>>
>> Sorry for my bad english.
>>
> My first thougth is that counter gets full and wrapped around more
> than one time between checks but then if mrtg gets it right... unless
> mrtg have some way of detecting how much counter overflows was between
> checks
> 
> Regards
> Mariusz

I'm not expert in snmp, however I know, snmp version 2c supports 64 bit 
counter instead 32 bit. And mrtg use it when I insert in config ::::2.
But I don't understand how I can use it in collectd.



-- 
С Уважением, Янченко Игорь
email: mailto:igor at skif.net.ua
jabber: xmpp://igor@jabber.ig0r.org.ua
SKIF ISP
IGR0-UANIC