Packet Discards

[ChesTrulson]

Hi,

It might also be an idea if you can trace the root cause of the problem.  Why is the traffic over running the switch buffers? 

It sounds like you might have devices that are unable to receive traffic at the required senders rate?

CheerZ

In my situation the ports that are dropping frames are in an MLT to a blade chassis switch.  That blade switch is feeding 16 servers, 4 of which are heavily loaded Novell proxy servers.  It could well be the case that blade switch (HP e2c) is not keeping up?  Proxy server traffic i can imagine would be very bursty too, which would make things worse for the 5530.

I'm not too sure what i can do to diagnose any further...

In future i'll be connecting blade chassis switches direct to 8600s wherever i can!

[jfarinha]

Looking on some advice on the same issue.

In my case, I have around 10.000 indiscards per hour (regular 3/second) on a few access ports that go to servers enclosures.

The ports are marked to "discard unregistered frames", but I wasn't expecting the servers to be sending tagged frames at all.

Any suggestion?

[Michael McNamara]

Looking on some advice on the same issue.

In my case, I have around 10.000 indiscards per hour (regular 3/second) on a few access ports that go to servers enclosures.

The ports are marked to "discard unregistered frames", but I wasn't expecting the servers to be sending tagged frames at all.

I would review the config of your servers, a packet trace of the traffic will probably answer the question.

Is there any non-IP traffic coming from the server(s)?

Good Luck!

[jfarinha]

Thanks! I will have a look and post any conclusions!

[bylie]

Is there any non-IP traffic coming from the server(s)?

Hmm, today I've actually also been busy looking at some inDiscard counters going up on one of our ERS 2500's uplinks. after verifying the usual VLAN tagging on both sides everything seems to be correct but still the counters keep on going up. I even tried disabling the "filter-untagged-frames" and "filter-unregistered-frames" features which we, by default, always enable on our uplinks, still no change. When I connected a laptop on the link to do a packetcapture nothing seemed out of the ordinary to me, but when reading your answer I recalled that I also saw some IPX traffic coming in (probably from some old equipment still thinking it's 1995 . Could this be the cause of this? I always asumed that a switch doesn't really care about which upper network protocol is used inside the frame (unless specific features are enabled ofcourse).
Is there actually any detailed explanation somewhere on what causes these Discard counters to go up? Aside from the usual mismatched VLAN tagging in combination with features like "filter-untagged-frames" and "filter-unregistered-frames"?

[Michael McNamara]

A high number of inDiscards could indicate a buffer resource problem (or put another way a traffic issue on the network).

Cheers!