Showing results for 
Search instead for 
Did you mean: 

Ask The Expert:QoS on Catalyst Switches

Community Manager
Community Manager

Read the bioWith Shashank Singh

Welcome to the Cisco Support Community Ask the Expert conversation. This is an opportunity to get an update on QoS on Catalyst 2960, 3550, 3560, 3750, 4500 and 6500 series switches with Cisco expert Shashank Singh. Shashank graduated in 2009 with a bachelor's degree in Computer Science and Engineering from VIT University, Vellore India. Prior to joining Cisco he worked at General Electric as a software engineer. Later on he joined the Cisco Technical Assistance Center as an engineer in October of 2009. He has been working on LAN Switching technologies in TAC since then. Shashank also holds a CCNP certificate. QoS on Catalyst switches is one of the areas of his interest.

Remember to use the rating system to let Shashank know if you have received an adequate response.

Shashank might not be able to answer each question due to the volume expected during this event. Remember that you can continue the conversation on the Network InfrastructureLAN Switching discussion forum shortly after the event. This event lasts through August 26 , 2011. Visit this forum often to view responses to your questions and the questions of other community members.

73 Replies 73

Hi Shashank,

I see:

Switch(config)#int gi1/0/45
Switch(config-if)#auto qos voip trust

*Mar  1 00:05:52.221: auto qos srnd4
*Mar  1 00:05:52.258:   mls qos trust cos
*Mar  1 00:05:52.274:  no queue-set 1
*Mar  1 00:05:52.279:   queue-set 1
*Mar  1 00:05:52.284:  priority-queue out
*Mar  1 00:05:52.284:  srr-queue bandwidth share 1 30 35 5

that's on a switch stack

Switch Ports Model              SW Version            SW Image
------ ----- -----              ----------            ----------
     1 52    WS-C2960S-48FPS-L  15.0(1)SE             C2960S-UNIVERSALK9-M
*    2 52    WS-C2960S-48FPS-L  15.0(1)SE             C2960S-UNIVERSALK9-M

Any idea what might be causing the difference between the queue-set commands that we see?


Lee Flight

Hi Lee,

I just confirmed that the behavior has been changed 12.2(58)SE onwards. IOS versions older than this one generate queue-set 2 while newer ones including 12.2(58)SE do not.



Hi Shashank,

thank for resolving this; perhaps you could pass this information back to the owners of the SBA guides for update in

any later revisions of those documents?

A QoS on Catalyst Ask the Expert would be a very useful event to have scheduled again (regularly).

Thank you for your diligence in reproducing this issue and your activity in this forum.

Lee Flight


I am still investigating whether or not the priority-queue, without congestion, actually will "overide" FIFO and forward the priority traffic before the other queues?

Hi Douglas,

At least on catalyst switches, this is the behavior. Priority queue is serviced until empty, and the link does not need to be congested for this to happen.




Thanks. Also what show commands can you use to see this?

Hi Douglas,

You can use "sh mls qos int gix/y statistics" to check the markings getting enqueued/dropped  on all the four queues.

Note that queue-1 will always be the priority queue on 3750/2960 switches once you configure 'priority-queue out'.  To check whether priority queue is enabled or not you may use the following command.

3700#sh mls qos  int gi3/0/1 queueing


Egress Priority Queue : enabled

Shaped queue weights (absolute) :  25 0 0 0

Shared queue weights  :  1 30 35 5

The port bandwidth limit : 100  (Operational Bandwidth:100.0)

The port is mapped to qset : 1



Hello Shashank,

Might sound a silly question, but I couldn't find a definitive answer for it.

What is considered congestion? Does it mean 1 packet sitting in the queue? 100% bandwidth being utilized?


Bruno Silva.

Hi Bruno,

"Conceptually, congestion is defined by the Cisco IOS software configuration guide as: "During periods of transmit congestion at the outgoing interface, packets arrive faster than the interface can send them."

In other words, congestion typically occurs when a fast ingress interface feeds a relatively slow egress interface.

Functionally, congestion is defined as filling the transmit ring on the interface. A ring is a special buffer control structure. Every interface supports a pair of rings: a receive ring for receiving packets and a transmit ring for transmitting packets."

Taken from

Hope that helps,




Have you come across the problem when you apply QoS queing for the 10G linecards in 6509 (WS-6704-10GE CFC and WS-6708-10GE DF3C) it prevents packet bytes greater than 718 bytes?

Sample queue:

interface TenGigabitEthernet1/1

no ip address

logging event link-status

load-interval 30

udld port aggressive

wrr-queue cos-map 1 1 2

wrr-queue cos-map 2 1 3

wrr-queue cos-map 3 1 1 6

wrr-queue cos-map 4 1 0 7

priority-queue cos-map 1 4 5

rcv-queue cos-map 1 1 4 5

rcv-queue cos-map 1 2 1 2 6

rcv-queue cos-map 1 3 3

rcv-queue cos-map 1 4 0 7

mls qos trust dscp


switchport trunk encapsulation dot1q

switchport trunk native vlan 500

switchport trunk allowed vlan 200,500

switchport mode trunk

Turning it on one side or both the  10G trunks is having this problem.

IRTest6509-01#debug qm error

QM error debugging is on

IRTest6509-01#debug qm error event port-err

IRTest6509-01#debug qm port-error

Port ASIC QoS error debugging is on

IRTest6509-01#debug qm port-error event

Port ASIC QoS event debugging is on

IRTest6509-01#debug qm port-eventint

QM interface events debugging is on


IRTest6509-01#conf t

Enter configuration commands, one per line.  End with CNTL/Z.





IRTest6509-01#ping si 10000

Type escape sequence to abort.

Sending 5, 10000-byte ICMP Echos to, timeout is 2 seconds:


Success rate is 100 percent (5/5), round-trip min/avg/max = 8/11/20 ms

IRTest6509-01#conf t

Enter configuration commands, one per line.  End with CNTL/Z.

IRTest6509-01(config)#mls qos


02:23:33:  HWIF QOS:hwif_qos_set_portqos_state: enqueueing port_qos_enable = 1, earl_qos_enable = 1

02:23:33: qos_enable_maccos(1): failed

02:23:33: QM enqueue global state change event for event:6, slot 0, up ENABLED

IRTest6509-01(config)#do ping si 718

Type escape sequence to abort.

Sending 5, 718-byte ICMP Echos to, timeout is 2 seconds:


Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/4 ms

IRTest6509-01(config)#do ping si 719

Type escape sequence to abort.

Sending 5, 719-byte ICMP Echos to, timeout is 2 seconds:


Success rate is 0 percent (0/5)

IRTest6509-01(config)#no mls qos


02:24:39:  HWIF QOS:hwif_qos_set_portqos_state: enqueueing port_qos_enable = 0, earl_qos_enable = 0

02:24:39: qos_enable_prec2cos(0): failed

02:24:39: qos_enable_maccos(0): failed

02:24:39: QM enqueue global state change event for event:6, slot 0, up DISABLED

02:24:40: hwif_qos_cos_mut_map_qos_statechange: TenGigabitEthernet1/1

02:24:40: hwif_qos_cos_mut_set_in_hardware: TenGigabitEthernet1/1

02:24:40: hwif_qos_cos_mut_set_in_hardware: Setting cos-mutation for TenGigabitEthernet1/1

IRTest6509-01(config)#do ping si 10000

Type escape sequence to abort.

Sending 5, 10000-byte ICMP Echos to, timeout is 2 seconds:


Success rate is 100 percent (5/5), round-trip min/avg/max = 8/10/12 ms


Would appreciate it.



Hi Erwyn,

I am not familiar with this issue. Please open a TAC case for further analysis.



Thanks, we have. Just trying to see if someone have already come across this, as there is no specific TAC case mentioned around it neither on the IOS bugs.




100mbps link with QoS vs 100mbps link without QoS

Both link are less than 50% utilized.

Both link have same source and same destination

Router is a decent model/router with 10% cpu utilization

Traffic is voice/jitter


Really a simple question:

Will there be a difference?



The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.

Liability Disclaimer

In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.


"It depends" on the nature of your traffic.

50% utilization tells us almost nothing.  You might desperately need QoS at 1% utilization or not need it at 100% utilization.

QoS is useful to manage congestion, including transient congestion, that's adverse to the service needs of your traffic.  This also assumes, such QoS congestion management can actually help.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Recognize Your Peers