From: Andreas Schlitter (andreasschlitter@yahoo.com)
Date: Mon Jul 28 2008 - 09:15:57 ART
Could this problem lie with the "sharing" feature of non-blocking ports on the 4500 ?
This "sharing" feature seems to work like a priority queue in that it reserves the bandwidth even if it does not use it.
I'm running a Sup2+ (WS-X4013+) and most (nearly 100%) of the traffic is flowing through the Tx-Queue-1 (by design).
4500#show qos interface gigabitEthernet 1/1
QoS is enabled globally
Port QoS is enabled
Administrative Port Trust State: 'dscp'
Operational Port Trust State: 'dscp'
Trust device: none
Default DSCP: 0 Default CoS: 0
Tx-Queue Bandwidth ShapeRate Priority QueueSize
(bps) (bps) (packets)
1 250000000 disabled N/A 1920
2 250000000 disabled N/A 1920
3 330000000 330000000 high 1920
4 250000000 disabled N/A 1920
4500#
4500#show interfaces gigabitEthernet 1/2 counters detail
Port Tx-Bytes-Queue-1 Tx-Bytes-Queue-2 Tx-Bytes-Queue-3 Tx-Bytes-Queue-4
Gi1/2 5499520174 0 2368 8286317
4500#
The way I see it a possible solution around this "sharing" feature would be to increase the Tx-Queue 1 from 250Mbps to 750Mbps.
Interface G1/1
tx-queue 1
bandwidth 750000000
Any comments ?
Thomas
----- Original Message ----
From: Andreas Schlitter <andreasschlitter@yahoo.com>
To: Muhammad Nasim <muhammad.nasim@gmail.com>
Cc: ccielab@groupstudy.com
Sent: Monday, July 28, 2008 11:06:01 AM
Subject: Re: AutoQoS & performance problem
Yes, of course.
Both devices directly connected to the 4500 as well as access switches connected from this switch have this issue.
Thanks, Andreas
----- Original Message ----
From: Muhammad Nasim <muhammad.nasim@gmail.com>
To: Andreas Schlitter <andreasschlitter@yahoo.com>
Cc: ccielab@groupstudy.com
Sent: Monday, July 28, 2008 10:11:47 AM
Subject: Re: AutoQoS & performance problem
Did you enable auto qos on access-layer switches where ip phones are going to connect or already connected?
2008/7/28 Andreas Schlitter <andreasschlitter@yahoo.com>
After configuring AutoQos (voip trust) on a network in order to support a small number of IP Telephones we are having problems with network performance.
Throughput of data traffic behind a Catalyst 4500 has droped to 1/10 of what is was before.
The network is not very busy and the voice traffic is not yet in production, so that isn't the problem.
This only seems to affect users connected through a Catalyst 4500, all other switches have normal performance.
Does anyone know what could be causing this or how I go about troubleshooting this ?
Thanks, Andreas
This archive was generated by hypermail 2.1.4 : Mon Aug 04 2008 - 06:11:57 ART