Lack of IPv6 traffic stats makes judging progress difficult

[Frank Bulk]

When I walk my 7600 the only interfaces that have values are VLANs.  My
guess is that these are only the software counters, not hardware?

Frank

-----Original Message-----
From: ipv6-ops-bounces+frnkblk=iname.com at lists.cluenet.de
[mailto:ipv6-ops-bounces+frnkblk=iname.com at lists.cluenet.de] On Behalf Of
Tassos Chatzithomaoglou
Sent: Wednesday, May 25, 2011 12:44 AM
To: ipv6-ops at lists.cluenet.de
Subject: Re: Lack of IPv6 traffic stats makes judging progress difficult

ok, i managed to ipv6 stats per interface on a CRS by
queryingipIfStatsHCIn/OutOctets:

i.e.
snmpwalk -v2c -c xxx router .1.3.6.1.2.1.4.31.3.1.6
ip.31.3.1.6.2.75 = Counter64: 193498077
ip.31.3.1.6.2.82 = Counter64: 1372480387
ip.31.3.1.6.2.85 = Counter64: 1026814419
ip.31.3.1.6.2.88 = Counter64: 30528758
ip.31.3.1.6.2.89 = Counter64: 1319486
ip.31.3.1.6.2.90 = Counter64: 278863122
ip.31.3.1.6.2.91 = Counter64: 61462823


7200 with latest SRE also returns those.

IPv4 stats from the same MIB are either not available or zero, so we'll 
stick with ifHCIn/OutOctets for those.

--
Tassos


Tassos Chatzithomaoglou wrote on 24/05/2011 23:32:
> Anyone tried ipIfStatsIPVersion under IP-MIB?
>
> That (in combination with other oids) should do the job, but i haven't 
> found a single platform that supports it, although CCO says otherwise.
>
> -- 
> Tassos
>
>
> Gert Doering wrote on 24/05/2011 23:00:
>> Hi,
>>
>> On Tue, May 24, 2011 at 03:16:09PM -0400, chip wrote:
>>> As a note, if you're collecting traffic on 6500/7600's you can get
>>> this by running the 'show interfaces accounting' command.  But you can
>>> only use the outbound counters, the inbound counters do not include
>>> hardware switched packets.
>> Actually, it's worse - hardware switched IPv6 packets are counted on
>> the IP*v4* counters.
>>
>> Cisco says it's a chip limitation...
>>
>> (We noticed this when we had a backbone link that only transitted IPv6
>> packets, but the IPv4 counters increased happily... opened a TAC case,
>> TAC could reproduce it, engineering came back with "wontfix" and so
>> the documentation was adjusted...)
>>
>> Gert Doering
>>          -- NetMaster
>