cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
4269
Views
0
Helpful
7
Replies

Qos on WAN Router & default class Offered rate

patelvc7601
Level 1
Level 1

I have question regarding Qos configuration on my WAN router. we have 100mbps WAN connection that feeds to mutliple site & we have applied Nested MQC. We are using vmwware view & that requires a lot of bandwidth so after voice & view have most of the bandwidth(60Mbps) is assigned .

After looking into the sh policy-map int gi0/1 command , I am finding that there are some drops on vmware view class & default class as well.

if you look at the config I have 5Mbps for voice & 60Mpbs for View traffic assigned. Before adding some more site I wanted to find out about these drops. I believe unclassified traffic will put into the default class but I would like to understand how this 30 second offered rate works ?

so if I have 30Mbps default class traffic than what will be my 30 secod offered rate ??( 30*30Mbps ) ??

Or View Traffic which is 60mbps guranted what will be my 30 second offered rate ? ( I can't believe that I am seeing drops on these)

Policy Map Metro-E
    Class class-default
      Average Rate Traffic Shaping
      cir 100000000 (bps)
      service-policy WAN-EDGE-ISP

Policy Map WAN-EDGE-ISP
    Class VOICE
      priority 5000
    Class VOICE-CTRL
      bandwidth 512
    Class VIDEO
      bandwidth 1000
    Class VMWARE_VIEW
    
bandwidth 60000
    Class class-default

      fair-queue

interface GigabitEthernet0/1
description ISP
ip address X.X.X.X 255.255.255.252
  duplex auto
speed auto
service-policy output Metro-E

==================================================

CHKD-WAS-COXATM-RTR#sh policy-map int gi0/1                
GigabitEthernet0/1

  Service-policy output: Metro-E

    Class-map: class-default (match-any)
      3135457351 packets, 861119166697 bytes
      30 second offered rate 5725000 bps, drop rate 16000 bps
      Match: any
      Queueing
      queue limit 64 packets
      (queue depth/total drops/no-buffer drops) 0/3247472/1980
      (pkts output/bytes output) 3134300024/864967258106
      shape (average) cir 100000000, bc 400000, be 400000
      target shape rate 100000000

      Service-policy : WAN-EDGE-ISP

        queue stats for all priority classes:
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 0/254/254
          (pkts output/bytes output) 479284491/82268615766

        Class-map: VOICE (match-any)
          479281966 packets, 75279196865 bytes
          30 second offered rate 313000 bps, drop rate 0 bps
          Match: ip dscp ef (46)
            9871531 packets, 1293626248 bytes
            30 second rate 18000 bps
          Match: ip rtp 16384 16383
            469413150 packets, 73985973595 bytes
            30 second rate 295000 bps
          Priority: 5000 kbps, burst bytes 125000, b/w exceed drops: 254

        Class-map: VOICE-CTRL (match-any)
          19186898 packets, 1737982174 bytes
          30 second offered rate 4000 bps, drop rate 0 bps
          Match: ip dscp af31 (26)
            11 packets, 1914 bytes
            30 second rate 0 bps
          Match: ip dscp cs3 (24)
            17192863 packets, 1225528998 bytes
            30 second rate 3000 bps
          Match: access-group name QoS_CallControl
            1994024 packets, 512451262 bytes
            30 second rate 0 bps
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 0/326/2
          (pkts output/bytes output) 19186576/1737855104
          bandwidth 512 kbps

        Class-map: VIDEO (match-any)
          0 packets, 0 bytes
          30 second offered rate 0 bps, drop rate 0 bps
          Match: ip dscp af41 (34)
            0 packets, 0 bytes
            30 second rate 0 bps
          Match: access-group name QOS_Video
            0 packets, 0 bytes
            30 second rate 0 bps
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 0/0/0
          (pkts output/bytes output) 0/0
          bandwidth 1000 kbps
         
        Class-map: VMWARE_VIEW (match-any)
          689809393 packets, 243968733796 bytes
         30 second offered rate 784000 bps, drop rate 0 bps
          Match: access-group name VMWARE_VIEW
            689809393 packets, 243968734588 bytes
            30 second rate 784000 bps
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 0/580301/619
          (pkts output/bytes output) 689229090/243609123774
          bandwidth 60000 kbps
         
        Class-map: class-default (match-any)
          1947181335 packets, 540134080193 bytes
          30 second offered rate 4620000 bps, drop rate 16000 bps
          Match: any
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops/flowdrops) 0/2666591/1105/2665486
          (pkts output/bytes output) 1946599864/537351661254
          Fair-queue: per-flow queue limit 16

Sincerely,

Viral

1 Accepted Solution

Accepted Solutions

Disclaimer

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.

Posting

re: "too fast"

Also another possible issue, many IOS shapers, don't account for L2 overhead.  I.e. your shaping for 100  Mbps could be "too fast".)

If this doesn't make sense, let me know.

re: "I would like to know how you came up with anywhere near 100 Mbps throughput (about 6  Mbps!)? "

From your 1st post stats, saw:

"30 second offered rate 5725000 bps, drop rate 16000 bps"

5725000 bps = about 6 Mbps, and its dropping packets.

I'm sure you do get more, I wasn't clear (sorry) - packets being dropped at such a low percentage, too me, implies a problem.

What's does the interface show for outbound bandwidth in comparison to the policy stats?

Still suspect your queues are too shallow for 100 Mbps.  What's your end-to-end, across WAN, latency?

View solution in original post

7 Replies 7

Joseph W. Doherty
Hall of Fame
Hall of Fame

Disclaimer

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.

Posting

BTW, from what you described, and before getting to your specific questions, although you're shaping for your logical WAN interface, what's the bandwidth to the other sites?  If not also 100 Mbps or if they are mutiple sites which can inter-communicate, you're likely to have congestion at MetroE egress.  (Also another possible issue, many IOS shapers, don't account for L2 overhead.  I.e. your shaping for 100 Mbps could be "too fast".)

The bandwidth statements, in non-LLQ classes, are used to set scheduling weight in such a way that you minimum bandwidth should be provided.  However, in IOSs w/o HQF like support, class-default can generate multiple flows each contending for bandwidth, which, I believe, can distort defined class bandwidths.

"Offered rate", I believe is a measure of the bandwidth the class "sees". It minus drop rate would be class transmission rate.

From your stats, since your not getting anywhere near 100 Mbps throughput (about 6 Mbps!?), and class-default is dropping packets(!), suspect indivdual queue depth, at 16, is much, much too shallow.  Overall queue depth, 64, might also be too shallow for 100 Mbps.  Some of the later IOSs allow configuration modifications of these.

Consider at 100 Mbps, 50 ms would be 5,000,000 bits or 625,000 bytes, or about 428 packets with 1460 payload.

BTW, from what you described, and before getting to your specific  questions, although you're shaping for your logical WAN interface,  what's the bandwidth to the other sites?  If not also 100 Mbps or if  they are mutiple sites which can inter-communicate, you're likely to  have congestion at MetroE egress.  (Also another possible issue, many  IOS shapers, don't account for L2 overhead.  I.e. your shaping for 100  Mbps could be "too fast".)

The bandwidth statements, in non-LLQ classes, are  used to set scheduling weight in such a way that you minimum bandwidth  should be provided.  However, in IOSs w/o HQF like support,  class-default can generate multiple flows each contending for bandwidth,  which, I believe, can distort defined class bandwidths.

"Offered  rate", I believe is a measure of the bandwidth the class "sees". It  minus drop rate would be class transmission rate.

From your  stats, since your not getting anywhere near 100 Mbps throughput (about 6  Mbps!?), and class-default is dropping packets(!), suspect indivdual  queue depth, at 16, is much, much too shallow.  Overall queue depth, 64,  might also be too shallow for 100 Mbps.  Some of the later IOSs allow  configuration modifications of these.

Consider at 100 Mbps, 50 ms would be  5,000,000 bits or 625,000 bytes, or about 428 packets with 1460  payload.

=============================================================================================

First of all thanks for your reply.

I have Hub & Spoke kind of topology & all my offsite traffic comes into WAN Router.  I am wondering if you can explain   "your shaping for 100  Mbps could be "too fast"

I would like to know how you came up with anywhere near 100 Mbps throughput (about 6  Mbps!)?

I am currenlty using Version 12.4(20)T on 3845 Series router.

I have sampled the stats : & trying to figure out how the my 30 second offfered rate is just 3.2Mbps & I still have drops on View Traffic.

WAN-RTR#sh policy-map int gi0/1 | be VMWARE

        Class-map: VMWARE_VIEW (match-any)
          8962790 packets, 3228673040 bytes
         30 second offered rate 3272000 bps, drop rate 19000 bps
          Match: access-group name VMWARE_VIEW
            8962790 packets, 3228673040 bytes
            30 second rate 3272000 bps
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 1/13133/0
          (pkts output/bytes output) 8949658/3220409068
          bandwidth 60000 kbps

        Class-map: class-default (match-any)
          22528969 packets, 7336523898 bytes
          30 second offered rate 5566000 bps, drop rate 30000 bps
          Match: any
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops/flowdrops) 8/68232/0/68232
          (pkts output/bytes output) 22468376/7260044601
          Fair-queue: per-flow queue limit 16

===============================================================================
WAN-RTR#sh policy-map int gi0/1 | be VMWARE
        Class-map: VMWARE_VIEW (match-any)
          8964413 packets, 3229562889 bytes
          30 second offered rate 3272000 bps, drop rate 19000 bps
          Match: access-group name VMWARE_VIEW
            8964413 packets, 3229562889 bytes
            30 second rate 3272000 bps (3.2Mbps)
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops) 0/13272/0
          (pkts output/bytes output) 8951141/3221185261
         bandwidth 60000 kbps

        Class-map: class-default (match-any)
          22531222 packets, 7337179674 bytes
          30 second offered rate 5566000 bps, drop rate 30000 bps
          Match: any
          Queueing
          queue limit 64 packets
          (queue depth/total drops/no-buffer drops/flowdrops) 0/68234/0/68234
          (pkts output/bytes output) 22470636/7260701639
          Fair-queue: per-flow queue limit 16

Sincerely,

Viral

Disclaimer

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.

Posting

re: "too fast"

Also another possible issue, many IOS shapers, don't account for L2 overhead.  I.e. your shaping for 100  Mbps could be "too fast".)

If this doesn't make sense, let me know.

re: "I would like to know how you came up with anywhere near 100 Mbps throughput (about 6  Mbps!)? "

From your 1st post stats, saw:

"30 second offered rate 5725000 bps, drop rate 16000 bps"

5725000 bps = about 6 Mbps, and its dropping packets.

I'm sure you do get more, I wasn't clear (sorry) - packets being dropped at such a low percentage, too me, implies a problem.

What's does the interface show for outbound bandwidth in comparison to the policy stats?

Still suspect your queues are too shallow for 100 Mbps.  What's your end-to-end, across WAN, latency?

Once Again Thank you for your reply. I removed the shaping but I was still getting drops on View traffic. I removed the Qos altogether :

For 5 hours & some minute no output drops. I am confused now.

sh int gi0/1

GigabitEthernet0/1 is up, line protocol is up

  Hardware is BCM1125 Internal MAC, address is 001e.be10.19c1 (bia 001e.be10.19c1)

  Description: COX-METRO-E PRINCESS/VOLVO

  Internet address is 10.21.251.1/24

  MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,

     reliability 255/255, txload 29/255, rxload 14/255

  Encapsulation ARPA, loopback not set

  Keepalive set (10 sec)

  Full-duplex, 100Mb/s, media type is RJ45

  output flow-control is XON, input flow-control is XON

  ARP type: ARPA, ARP Timeout 04:00:00

  Last input 00:00:00, output 00:00:00, output hang never

  Last clearing of "show interface" counters 00:03:47

  Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

  Queueing strategy: fifo

  Output queue: 77/2000 (size/max)

  30 second input rate 5660000 bits/sec, 2367 packets/sec

  30 second output rate 11692000 bits/sec, 2320 packets/sec

     454942 packets input, 174325805 bytes, 0 no buffer

     Received 211 broadcasts, 0 runts, 0 giants, 0 throttles

     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

     0 watchdog, 9279 multicast, 0 pause input

     0 input packets with dribble condition detected

     396532 packets output, 178155525 bytes, 0 underruns

     0 output errors, 0 collisions, 0 interface resets

     0 unknown protocol drops

     0 babbles, 0 late collision, 0 deferred

     0 lost carrier, 0 no carrier, 0 pause output

     0 output buffer failures, 0 output buffers swapped out

GigabitEthernet0/1 is up, line protocol is up
  Hardware is BCM1125 Internal MAC, address is 001e.be10.19c1 (bia 001e.be10.19c1)
  Description: COX-METRO-E PRINCESS/VOLVO
  Internet address is 10.21.251.1/24
  MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,
     reliability 255/255, txload 22/255, rxload 21/255
  Encapsulation ARPA, loopback not set
  Keepalive set (10 sec)
  Full-duplex, 100Mb/s, media type is RJ45
  output flow-control is XON, input flow-control is XON
  ARP type: ARPA, ARP Timeout 04:00:00
  Last input 00:00:03, output 00:00:00, output hang never
  Last clearing of "show interface" counters 05:26:24
  Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
  Queueing strategy: fifo
  Output queue: 0/2000 (size/max)
  30 second input rate 8326000 bits/sec, 3433 packets/sec
  30 second output rate 8865000 bits/sec, 2962 packets/sec
     76084253 packets input, 164780924 bytes, 0 no buffer
     Received 34671 broadcasts, 0 runts, 0 giants, 0 throttles
     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
     0 watchdog, 1130392 multicast, 0 pause input
     0 input packets with dribble condition detected
     62475429 packets output, 1028058297 bytes, 0 underruns
     0 output errors, 0 collisions, 0 interface resets
     86 unknown protocol drops
     0 babbles, 0 late collision, 0 deferred
     0 lost carrier, 0 no carrier, 0 pause output
     0 output buffer failures, 0 output buffers swapped out

I hope somebody can explain me 30 second offered rate & drop rate ( what should be this value in ideal word if  I have assigned 60Mbps to view traffic)

sh policy-map int gi0/1 | be VMWARE_VIEW

    Class-map: VMWARE_VIEW (match-any)

      158755 packets, 54247240 bytes

      30 second offered rate 2351000 bps, drop rate 0 bps

      Match: access-group name VMWARE_VIEW

        158755 packets, 54247240 bytes

        30 second rate 2351000 bps

      Queueing

      queue limit 128 packets

      (queue depth/total drops/no-buffer drops) 0/6/0

      (pkts output/bytes output) 158747/54243288

      bandwidth 60000 kbps

    Class-map: class-default (match-any)

      592868 packets, 210011906 bytes

      30 second offered rate 7422000 bps, drop rate 105000 bps

      Match: any

      Queueing

      queue limit 64 packets

      (queue depth/total drops/no-buffer drops/flowdrops) 3/1609/0/1609

      (pkts output/bytes output) 591574/207979753

      Fair-queue: per-flow queue limit 16

############################################################################

sh policy-map int gi0/1 | be VMWARE_VIEW

    Class-map: VMWARE_VIEW (match-any)

      159579 packets, 54614680 bytes

      30 second offered rate 2533000 bps, drop rate 2000 bps ( at one point I was thinking that 2351000 bps is the value ) But other output did not match but I still do not know how IOS is coming up with this 30 second number offered rate & drop rate )

      Match: access-group name VMWARE_VIEW

       159579 packets, 54614680 bytes

        30 second rate 2533000 bps

      Queueing

      queue limit 128 packets

      (queue depth/total drops/no-buffer drops) 4/6/0

      (pkts output/bytes output) 159574/54611606

      bandwidth 60000 kbps

    Class-map: class-default (match-any)

      595177 packets, 211237940 bytes

      30 second offered rate 7086000 bps, drop rate 91000 bps

      Match: any

      Queueing

      queue limit 64 packets

      (queue depth/total drops/no-buffer drops/flowdrops) 43/1616/0/1616

      (pkts output/bytes output) 593878/209195765

      Fair-queue: per-flow queue limit 16

Disclaimer

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.

Posting

patelvc7601 wrote:

Once Again Thank you for your reply. I removed the shaping but I was still getting drops on View traffic. I removed the Qos altogether :

For 5 hours & some minute no output drops. I am confused now.

sh int gi0/1

GigabitEthernet0/1 is up, line protocol is up

  Hardware is BCM1125 Internal MAC, address is 001e.be10.19c1 (bia 001e.be10.19c1)

  Description: COX-METRO-E PRINCESS/VOLVO

  Internet address is 10.21.251.1/24

  MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,

     reliability 255/255, txload 29/255, rxload 14/255

  Encapsulation ARPA, loopback not set

  Keepalive set (10 sec)

  Full-duplex, 100Mb/s, media type is RJ45

  output flow-control is XON, input flow-control is XON

  ARP type: ARPA, ARP Timeout 04:00:00

  Last input 00:00:00, output 00:00:00, output hang never

  Last clearing of "show interface" counters 00:03:47

  Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

  Queueing strategy: fifo

  Output queue: 77/2000 (size/max)

  30 second input rate 5660000 bits/sec, 2367 packets/sec

  30 second output rate 11692000 bits/sec, 2320 packets/sec

     454942 packets input, 174325805 bytes, 0 no buffer

     Received 211 broadcasts, 0 runts, 0 giants, 0 throttles

     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

     0 watchdog, 9279 multicast, 0 pause input

     0 input packets with dribble condition detected

     396532 packets output, 178155525 bytes, 0 underruns

     0 output errors, 0 collisions, 0 interface resets

     0 unknown protocol drops

     0 babbles, 0 late collision, 0 deferred

     0 lost carrier, 0 no carrier, 0 pause output

     0 output buffer failures, 0 output buffers swapped out

Look at your queue limit (20000) and current queue depth (77).  Compare to prior QoS queue limits.

If Anybody have read this post, I wanted to point this out that all these drops turn out to be duplex mismatch with my WAN Carrier. I knew about duplex mismatch is bad but this was not direct connection it was indirect connection which caused the issue. I had a open case with Carrier for few month & one fine day somebody was able to figure out.

My WAN Carrier was providing me the Etherenet Link & I had to stretch out to another building because my wan router was another building . I used the media converter between my WAN Router & Carrier Equipment, Duplex mismatch was between Media Converter to my Carrier Equipment( I did not have any visibility in terms of Error etc..) . I could not see any input or output errors on my Router because my router to media converter connection was good.

In short topology is like :

WAN(R) gi0/1 coper>>>100Base T Media Con 100Base Fiber.<<<<>>>>100Base Fiber Media Con,100Base T>>>>Carrier Equipment

With the same config for week now , Everything looks good. I had a TAC Case opened on this & they told me that micro burst is giving me those output drops.

My output drops disappear.

GigabitEthernet0/1 is up, line protocol is up

  Hardware is BCM1125 Internal MAC, address is 001e.b311.19c1 (bia 001e.b311.19c1)

  Description:

  Internet address is x.x.x.x/24

  MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,

     reliability 255/255, txload 23/255, rxload 38/255

  Encapsulation ARPA, loopback not set

  Keepalive set (10 sec)

  Full Duplex, 100Mbps, media type is RJ45

  output flow-control is XON, input flow-control is XON

  ARP type: ARPA, ARP Timeout 04:00:00

  Last input 00:00:00, output 00:00:00, output hang never

  Last clearing of "show interface" counters 5d14h

  Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

  Queueing strategy: Class-based queueing

  Output queue: 0/2000/0 (size/max total/drops)

  30 second input rate 15284000 bits/sec, 6613 packets/sec

  30 second output rate 9138000 bits/sec, 5025 packets/sec

     1231779335 packets input, 1547946014 bytes, 0 no buffer

     Received 543988 broadcasts (274684 IP multicasts)

     0 runts, 0 giants, 0 throttles

     0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

     0 watchdog, 553599 multicast, 0 pause input

     911257186 packets output, 1770949254 bytes, 0 underruns

     0 output errors, 0 collisions, 0 interface resets

     0 unknown protocol drops

     0 babbles, 0 late collision, 0 deferred

     0 lost carrier, 0 no carrier, 0 pause output

     0 output buffer failures, 0 output buffers swapped out

Thank you

Viral

Disclaimer

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.

Posting

Transient congestion might have been caused by retransmissions.

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: