QOS betweeen 4510 and 3550

satish rawat · ‎11-01-2011

I am having issues working on my QOS between 4510 and 3550 switch connecting on layer 3 through a service provider.

I have class maps and policy map setup on both sides and then policy map attached to interfaces however i dont see any traffic matching in policy map on 3550 switch, i do see some traffic matching on 4510 but the speed with which its increasing has my doubts about it. When i make voip calls ( VOIP switches are sitting behind 3550 and are mainly 3550 pwr 24 port switches with phone ports configured for auto qos voip cisco-phone and trusting cos)

i rarely see the RTP matching in class under policy map.

Please see my config and let me know if anyone has any suggestions.

Thanks

acampbell · ‎11-01-2011

Hi,

From the 3550 config book

http://www.cisco.com/en/US/partner/docs/switches/lan/catalyst3550/software/release/12.2_44_se/configuration/guide/swqos.html#wp1048656

You cannot use the service-policy interface configuration command to attach policy maps that contain these elements to an egress interface:

This affects almost all 35xx,37zxx lan switches.

On a 3550 try using this:-

c3550#conf t
c3550(config)#int fas 0/1
c3550(config-if)#mls qos monitor packet
c3550(configif)#mls qos monitor dscp 26 46

c3550#show mls qos interface f0/1 statistics

HTH
Alex

Please rate useful posts

Regards, Alex. Please rate useful posts.

satish rawat · ‎11-01-2011

Alex,

Thanks for your quick response.

no changes on above config and Today i noticed some traffic being matched on policy map applied to 4510 layer 3 interface connecting to 3550, however no traffic on policy map on 3550 side.

i read somewhere that there is a known issue with 3550 switches where they dont show any traffic stats on policy map but what is stragne is default class does show traffic matched but not on classes created for RTP and SIG

4510#sh policy-map interface gigabitEthernet 9/2
GigabitEthernet9/2

Service-policy output: autoqos-voip-policy

    Class-map: VOICE-RTP (match-any)
      3762905 packets
      Match: ip dscp ef
        3762905 packets
      QoS Set
       ip dscp ef
      police: Per-interface
        Conform: 11359400 bytes Exceed: 0 bytes

    Class-map: VOICE-SIG (match-all)
      41544 packets
      Match: ip dscp cs3 af31
      Match: ip dscp af31
      QoS Set
       ip dscp cs3
      police: Per-interface
        Conform: 0 bytes Exceed: 0 bytes

    Class-map: class-default (match-any)
      426812434 packets
      Match: any
TRHS-MDF-4510#sh qos interface gigabitEthernet 9/2
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        31250000    disabled    N/A        292
2        31250000    disabled    N/A        292
3        33000000    33000000    high       292
4        31250000    disabled    N/A        292

below is output from 3550

3550G-01#sh policy-map interface gigabitEthernet 0/12
GigabitEthernet0/12

Service-policy output: VOIP

    Class-map: VOICE-RTP (match-any)
      0 packets, 0 bytes
      5 minute offered rate 0 bps, drop rate 0 bps
      Match: ip dscp ef (46)
        0 packets, 0 bytes
        5 minute rate 0 bps

    Class-map: VOICE-SIG (match-any)
      0 packets, 0 bytes
      5 minute offered rate 0 bps, drop rate 0 bps
      Match: ip dscp cs3 (24) af31 (26)
        0 packets, 0 bytes
        5 minute rate 0 bps

    Class-map: class-default (match-any)
      61156 packets, 5431666 bytes
      5 minute offered rate 1000 bps, drop rate 0 bps
      Match: any
        61156 packets, 5431666 bytes
        5 minute rate 1000 bps

3550G-01#sh mls qos interface gigabitEthernet 0/12 st
GigabitEthernet0/12
Ingress
dscp: incoming   no_change classified policed    dropped (in bytes)
Others: 127200336 3861291969 560875663 0          0
Egress
dscp: incoming   no_change classified policed    dropped (in bytes)
Others: 870205991     n/a       n/a      0          0

WRED drop counts:
qid thresh1    thresh2   FreeQ
   1 : 0          0         2049
   2 : 0          0         1024
   3 : 0          0         614
   4 : 0          0         409

any ideas...

FYI : Alex i have not tried monitor commands you suggested but i am going to try.

another question : on 4510 policy map output for class RTP i see more confron packets than matched, any idea about why would that be a case?

thanks in advance

Kishore Chennupati · ‎11-01-2011

Hi satish,

3550G-01#sh mls qos interface gigabitEthernet 0/12 st

GigabitEthernet0/12
Ingress
dscp: incoming   no_change classified policed    dropped (in bytes)
Others: 127200336 3861291969 560875663 0          0
Egress
dscp: incoming   no_change classified policed    dropped (in bytes)
Others: 870205991     n/a       n/a      0          0

WRED drop counts:
qid thresh1    thresh2   FreeQ
   1 : 0          0         2049
   2 : 0          0         1024
   3 : 0          0         614
   4 : 0          0         409

As you can see from your output above. The incoming frames are being classified approprately but your outgoing ones are not. What happens when you use mls qos trust dscp instead of mls qos trust cos on interface with the VOIP phones on?

And also you can put the mls qos monitor command that Alex. Its give you a good idead as which ones are being marked etc.

another question : on 4510 policy map output for class RTP i see more confron packets than matched, any idea about why would that be a case?

conformed packets are in bytes whereas the matched are in packets

HTH

Kishore

satish rawat · ‎11-02-2011

This is what i am getting on running monitor command, very confusing stats

didnt tried changing trust dscp on phones yet, is that a good idea ?

3550G-01#sh mls qos interface gigabitEthernet 0/12 st
GigabitEthernet0/12
Ingress
dscp: incoming   no_change classified policed    dropped (in pkts)
    24: 0          0          0          0          0
    46: 6          0          0          0          0
Others: 143198     143138     66         0          0
Egress
dscp: incoming   no_change classified policed    dropped (in pkts)
    24: 2090          n/a       n/a      0          0
    46: 1             n/a       n/a      0          0
Others: 108692        n/a       n/a      0          0

WRED drop counts:
qid thresh1    thresh2   FreeQ
   1 : 0          0         2049
   2 : 0          0         1024
   3 : 0          0         614
   4 : 0          0         409

Kishore Chennupati · ‎11-02-2011

Satish,

Techinally , it should work with your cos values as you have changed the DSCP from 40 to 46 in your cos-dscp map.

How are your phones connected to the switch. Are they connected via just access ports or like a phone and PC connected back to back.

can you type the command sh mls qos interface statistics on the port where your phone is connected. This will show you whether the port is actually trusting them.

For eg:

sw1#sh mls qos interface fastEthernet 3/0/11 statistics

FastEthernet3/0/11 (All statistics are in packets)

dscp: incoming

-------------------------------

0 - 4 : 23368 0 0 0 0

5 - 9 : 0 0 0 0 0

10 - 14 : 0 0 0 0 0

15 - 19 : 0 0 0 0 0

20 - 24 : 0 0 0 0 0

25 - 29 : 0 0 0 0 0

30 - 34 : 0 0 0 0 16

35 - 39 : 0 0 0 0 0

40 - 44 : 0 0 0 0 0

45 - 49 : 0 3294965 0 0 0

I would try with mls qos trust dscp . Nothing wrong in testing one or 2 phones.It's not as if you are going to break something major. Put that config on 1 or 2 phones and make some phone calls and see if that helps.

Kishore