underrun errors on 3500xl

[n.titchener]

I have a problem!

We have about 85 3500XL switches in our network. several switches within the network show underrun errors on all fastethernet interfaces. For most of the time these errors increment slowly, about one every minute or so, but occasionally there will be a period of many errors over a short time.

The buffer settings on thse switches are default. The traffic throughout the network is similar and the load/throughput of all switche is also similar.

This issue is not causing problems at the moment but I think this may be a symptom of an underlying problem.

Any suggestions?

[owillins]

The general error percentage goal is to stay 1% of the total input/output packets. The underrun condition occurs when all egress ports are trying to access the exact same packet from the shared data buffer or the connecting devices are not able to accept the number of packets or the speed at which they are being sent. By splitting the switch into multiple VLANs, fewer egress ports will be accessing the same packet. If multiple VLANs per switch cannot be configured due to network design restrictions, limit the number of interfaces accessing the same multicast stream.

[n.titchener]

The number of underruns is above 1%. All ports on the three switches are in the same Vlan. However, most ports on only two of the switches are experiencing these underruns. Because only two switches are experiencing the underruns i think it's fair to say that they are not being caused by broadcasts and viewing ip traffic shows no multicasts received or sent. The numbers of underruns does vary from port to port between 3%-0.5%. As I mentioned it's not causing a problem at the moment, I'm interested in finding out if anyone knows of a reason why this is happening.

Thanks for your reply