Just in case there are any Cisco certified engineers here...

I'm trying to help debug a network issue:

Running a Cisco Catalyst 6509 with IOS software. After a lot of hassle, finally managed to get CiscoView 6 installed and running (as part of Ciscoworks) in order to try and fault find an issue we have where a networked system stops functioning. :(

Looking at the status of the Supervisor engine (it's a Sup720 ), I get a shed load of UDP errors, just about as many errors as UDP packets sent over the network :eek:

I've set up a span monitoring port on the cisco and have ethereal running on a host PC capturing all network traffic, but as you can imagine, there's a fair few million packets whizzing around the place and I don't know where to start

However, if someone was able to technically explain what is meant by a "UDP error", i.e. what does a UDP packet need to have / not have for it to be considered erroneous

That way, it might help me narrow down and find the cause of the issue.

Thanks

Oh, and Zoid, in case you're wondering, yes, it's that switch which required those special high-speed fan trays

Any way you could do a packet debug on udp and send it to me? The graphs mean nothing to me out of context.

packet debug :confused:

How does one do that

packet debug :confused:

How does one do that

Oh dear... If you ask that then I wouldn't do it :eek:

Go get a PC and download ethereal from the internet and turn it on on the swtich thats giving you grief. (That'll keep you away from the cisco side of things) Ethereal is a free packet sniffer and will give you an idea of whats happening on the switch. More than likely, if the "network stops working" then there's a loop somewhere. Check all cabling! (Yup I did say it! - Its gonna take hours if the 6509 is fully loaded)

LOL You haven't read my first post :fap: :P

I have set up monitoring to capture all traffic on the vlan using ethereal. Very handy tool too

I know what's kicking out all the UDP packets, I just want to know why / how they are considered erroneous. Is it just the sheer amount that makes the Cisco think there's an error and hence logs them as bad udp packets?

I've checked all those udp packets with ethereal and their CRCs are OK, so can't think of what other "bad" stuff could be.

Regarding loops, I doubt it. Although spanning-tree is enabled, we only have two switches linked together via 8 ports of etherchannel, and although some devices are dual-attached, they're configured as failover rather than load-balancing so I van't see how we have some loops in the system. Or were you referring to something else?

Thanks for the input so far

Doh! I should've re-read it!

What ports are the UDP packets using? It may give a indication. PM me some screenshots, I'll be able to diagnose further

If spanning tree is enabled then you should be fine (It is enabled on all ports?), I'm starting to think along the lines of a virus... (I wont be sure until I see some UDP packets from ethereal)