Interface overruns and WRR queuing

[rmessina]

I have a gigabit uplink between our core switch and top of rack switch with mulitple database servers connected.  Our monitoring software is showing that this interface is discarding around 1000 or more packets an hour.  I don't believe that the interface is configured properly and that the queue size is causing the packet overruns on the interface.  I do not know enough about wrr queuing to fully understand the config.  Any asistance or documentation on wrr would be greatly appreciated.  Thanks!

interface GigabitEthernet1/4

switchport

switchport trunk encapsulation dot1q

switchport mode trunk

priority-queue queue-limit 5

wrr-queue queue-limit 65 15 15

wrr-queue random-detect min-threshold 1 70 100 100 100 100 100 100 100

wrr-queue random-detect min-threshold 2 70 100 100 100 100 100 100 100

wrr-queue random-detect min-threshold 3 40 40 50 50 60 60 70 70

wrr-queue random-detect max-threshold 1 100 100 100 100 100 100 100 100

wrr-queue random-detect max-threshold 2 100 100 100 100 100 100 100 100

wrr-queue random-detect max-threshold 3 70 70 80 80 90 90 100 100

wrr-queue cos-map 2 1 1 2

wrr-queue cos-map 3 5 3 4

wrr-queue cos-map 3 7 6 7

mls qos trust cos

[sdheer]

Hi Randy,

Overruns are basically caused dut to the flow control done by the ports and the fabric.

The concept behind overruns is when the packets coming from any interface goes to the fabric at such high rate that fabric is not able to handle  them all at a time , it will send a flow control to the interafce asking it to stop sending the frames for the time being until it gets free to take in and process other frames , this  will cause overruns to increment .

Same flow control happens other way as well .

Please check the fabric utilisation using command " show fabric utilisation" .

In order to understand the qos config on teh switch , below is teh link.

http://www.cisco.com/en/US/docs/switches/lan/catalyst6500/ios/12.2SXF/native/configuration/guide/qos.html

http://www.cisco.com/en/US/products/hw/switches/ps5023/products_tech_note09186a0080883f9e.shtml

Swati

Please rate if the content was found useful

[rmessina]

A show interface indicates that flow control is off.  The QOS config was generated by auto QOS.  The show fabric command does not work on my 6500.  I believe I just need to increase the size of the queue for this interface given the amount of traffic coming in from the other side.

[rsimoni]

Hi Randy,

as Swati pointed out overruns on cat6500 are indication of drops occurring at fabric level. Tuning WRR won't help as you will only changing queueuing at port asic level while the drops are occurring somewhere else.

Interface flow control is not related to fabric flow control either.

It is very complicated to discuss about fabric related troubleshooting on the support forum as deep architecture understanding is needed.

If you account is entitled I suggest you to open a TAC case from this thread and add a show tech of your device as a first step. The TAC engineer will ask you a bunch of other outputs to troubleshoot the problem further.

take care,

Riccardo