[mrtg] cfgmaker snmp Drop back to V1 RESOLVED
Pavel Ruzicka
pavouk at pavouk.org
Sat Mar 22 10:50:41 CET 2008
Hello,
I resolved my problem with fallback.
I found reason, why cfgmaker doesn't generate NoHC[] directive.
It is because I use my own template with --if-template option.
I generate Target in template with:
----------------------------------------------
Target[$target_name]: $if_ref:$router_connect
----------------------------------------------
When I change this line to:
----------------------------------------------
$default_target_directive
----------------------------------------------
It now generates:
----------------------------------------------
Target[$target_name]: $if_ref:$router_connect
noHC[$target_name]: yes
----------------------------------------------
It completely resolve my problem.
This behaviour with "$default_target_directive" is not documented in cfgmaker
documentation. I found it after look in cfgmaker perl code. Probably example
Template in cfgmaker documentation need change too.
Please change somebody documentation, if you can.
Thank you very much. I am happy now :-)
Best regards,
Pavel Ruzicka
> > in cfgmaker 2.16.1 (2.15.2, 2.16.0) doesn't work
> > fallback to SNMP V1 if check for HighspeedCounters failed.
> >
> > It works perfectly in 2.15.1.
>
> In 2.15.2 the NoHC[] directive was added to solve problems with V3 and
> fallback. With snmpv3, you can't fall back to v1, as the credentials
> are completely different, so we needed a different signal to identify
> when to use the HC counter ans when not to.
>
> What cfgmaker should be doing is leaving the polling as snmpV2 or V3,
> but add a NoHC[] directive which should cause mrtg to poll ifInOctets
> and ifOutOctets instead of ifHCInOctets and ifHCOutOctets.
>
> I haven't looked at 2.16.0/1 yet, but see if cfgmaker is adding the
> NoHC[] directive. If it is, then it is a libmrtg problem that needs to
> be fixed. If not, then it is a cfgmaker problem that needs to be fixed.
More information about the mrtg
mailing list