Re: Version 2 sflow: source_id not matching input or output port

From: brandy3389 <brandy3389@yahoo.com>
Date: 01/25/07
Message-ID: <139687.30744.qm@web36312.mail.mud.yahoo.com>

Appreciate your response.
  I checked that the 'class' is always 0, when this situation occurs. I am replaying a file from a customer site, so I don't have access to the device sending the sflow packets. Do you have any idea if the index is bogus, and I should use the input_port? I'm tracking interface information by the port number.
   
  Thanks

Neil McKee <neil.mckee@inmon.com> wrote:
  The source-id comes in two parts: "class" and "index". The class is
usually 0, meaning that the index is an ifIndex number,
corresponding to an entry in the SNMP MIB-II interfaces table. Is
there an entry in the SNMP ifTable for ifIndex = 99? (try "snmpwalk
-v1 -c interfaces"). It may not be a physical
interface. For example, it might represent the interface that
connects to the internal CPU on the device. It could even represent
an ACL.

It's also possible that the "class" could be 1, which means that the
index is a vlan, or 2 which means the index is for entry in the
Entity MIB. However I don't know of any implementations that use
these options.

Hope this helps.

regards,
neil

On Jan 25, 2007, at 9:32 AM, brandy3389 wrote:

> I'm seeing sflow packets with the source id of 99, which doesn't
> match either the input port or the output port. Also, this appears
> to occur when there is extended switch data, but not everytime. Any
> ideas what may be causing this.
>
> Thanks
>
>
> ---------------------------------
> Food fight? Enjoy some healthy debate
> in the Yahoo! Answers Food & Drink Q&A.

----------
Neil McKee
http://www.inmon.com

 
---------------------------------
Check out the all-new Yahoo! Mail beta - Fire up a more powerful email and get things done faster.
Received on Thu Jan 25 11:21:05 2007

This archive was generated by hypermail 2.1.8 : 01/25/07 PST