[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[altq:1709] Re: PRIQ problem

To: "Nicolas Christin" <nicolas@cs.virginia.edu>
Subject: [altq:1709] Re: PRIQ problem
From: "newbie" <dynks@polbox.com>
Date: Thu, 14 Nov 2002 17:15:22 +0100
Cc: "ALTQ Mailing-List" <altq@csl.sony.co.jp>
References: <Pine.GSO.4.44.0211141015560.24226-100000@arachnion.cs.Virginia.EDU>
Sender: owner-altq@csl.sony.co.jp
> No. I'm talking about the buffer size of your output link. Could you
> give us the relevant part (i.e., when you observe it doesn't work the
> way you want it to work) of your altqstat output? An altqstat output
> of the case where you experience satisfactory experience would also be
> good, to compare the two cases.

Thanks for fast response :-)
---------------------------------------------
here is the present config :

interface rl0 bandwidth 0.10M tbrsize 500 priq
class priq rl0 outgoing_udp_packets NULL priority 3
class priq rl0 outgoing_ping NULL priority 2
class priq rl0 ssh NULL priority 1
class priq rl0 default_class NULL priority 0 default
filter rl0 outgoing_udp_packets 0 0 0 27005 17
filter rl0 outgoing_ping 0 0 0 0 1
filter rl0 ssh 0 0 0 22 6
------------------------------------------------

and here is the output from altqstat during huge ftp xfer :

rl0:
[outgoing_udp_packets] handle:0xc2446180 pri:3
  measured: 11.14Kbps qlen: 1 period:3166
     packets:4255 (365001 bytes) drops:0
[outgoing_ping] handle:0xc24429c0 pri:2
  measured: 0bps qlen: 0 period:176
     packets:176 (14728 bytes) drops:0
[ssh] handle:0xc244b700 pri:1
  measured: 1.89Kbps qlen: 0 period:443
     packets:502 (109533 bytes) drops:0
[default_class] handle:0xc2460a00 pri:0
  measured: 95.79Kbps qlen: 7 period:4155
     packets:5800 (1280685 bytes) drops:39
     RED q_avg:10.03 xmit:5807 (forced:0 early:39 marked:0)

rl0:
[outgoing_udp_packets] handle:0xc2446180 pri:3
  measured: 11.77Kbps qlen: 2 period:3177
     packets:4273 (366474 bytes) drops:0
[outgoing_ping] handle:0xc24429c0 pri:2
  measured: 0.56Kbps qlen: 0 period:177
     packets:177 (14798 bytes) drops:0
[ssh] handle:0xc244b700 pri:1
  measured: 6.92Kbps qlen: 0 period:444
     packets:505 (110399 bytes) drops:0
[default_class] handle:0xc2460a00 pri:0
  measured: 76.81Kbps qlen: 6 period:4155
     packets:5815 (1290295 bytes) drops:39
     RED q_avg:9.82 xmit:5821 (forced:0 early:39 marked:0)

rl0:
[outgoing_udp_packets] handle:0xc2446180 pri:3
  measured: 11.89Kbps qlen: 2 period:3185
     packets:4291 (367961 bytes) drops:0
[outgoing_ping] handle:0xc24429c0 pri:2
  measured: 0.56Kbps qlen: 0 period:178
     packets:178 (14868 bytes) drops:0
[ssh] handle:0xc244b700 pri:1
  measured: 6.92Kbps qlen: 0 period:445
     packets:508 (111265 bytes) drops:0
[default_class] handle:0xc2460a00 pri:0
  measured: 76.50Kbps qlen: 7 period:4155
     packets:5830 (1299866 bytes) drops:39
     RED q_avg:8.94 xmit:5837 (forced:0 early:39 marked:0)

rl0:
[outgoing_udp_packets] handle:0xc2446180 pri:3
  measured: 13.66Kbps qlen: 1 period:3195
     packets:4312 (369670 bytes) drops:0
[outgoing_ping] handle:0xc24429c0 pri:2
  measured: 0bps qlen: 0 period:178
     packets:178 (14868 bytes) drops:0
[ssh] handle:0xc244b700 pri:1
  measured: 6.92Kbps qlen: 0 period:446
     packets:511 (112131 bytes) drops:0
[default_class] handle:0xc2460a00 pri:0
  measured: 86.87Kbps qlen: 7 period:4155
     packets:5842 (1310734 bytes) drops:40
     RED q_avg:8.49 xmit:5849 (forced:0 early:40 marked:0)

rl0:
[outgoing_udp_packets] handle:0xc2446180 pri:3
  measured: 12.33Kbps qlen: 1 period:3206
     packets:4331 (371213 bytes) drops:0
[outgoing_ping] handle:0xc24429c0 pri:2
  measured: 0.56Kbps qlen: 0 period:179
     packets:179 (14938 bytes) drops:0
[ssh] handle:0xc244b700 pri:1
  measured: 6.79Kbps qlen: 0 period:447
     packets:514 (112981 bytes) drops:0
[default_class] handle:0xc2460a00 pri:0
  measured: 77.79Kbps qlen: 3 period:4155
     packets:5860 (1320466 bytes) drops:40
     RED q_avg:7.56 xmit:5863 (forced:0 early:40 marked:0)
------------------------------------------------------------

and here is how ping to the 1st hop looks during huge ftp transfer :

PING xxx.76.32.153 (xxx.76.32.153): 56 data bytes
64 bytes from xxx.76.32.153: icmp_seq=0 ttl=254 time=73.364 ms
64 bytes from xxx.76.32.153: icmp_seq=1 ttl=254 time=44.135 ms
64 bytes from xxx.76.32.153: icmp_seq=2 ttl=254 time=117.453 ms
64 bytes from xxx.76.32.153: icmp_seq=3 ttl=254 time=59.180 ms
64 bytes from xxx.76.32.153: icmp_seq=4 ttl=254 time=126.494 ms
64 bytes from xxx.76.32.153: icmp_seq=5 ttl=254 time=86.671 ms
64 bytes from xxx.76.32.153: icmp_seq=6 ttl=254 time=16.312 ms
64 bytes from xxx.76.32.153: icmp_seq=7 ttl=254 time=85.409 ms
64 bytes from xxx.76.32.153: icmp_seq=8 ttl=254 time=21.803 ms
64 bytes from xxx.76.32.153: icmp_seq=9 ttl=254 time=86.737 ms
64 bytes from xxx.76.32.153: icmp_seq=10 ttl=254 time=25.031 ms
64 bytes from xxx.76.32.153: icmp_seq=11 ttl=254 time=110.261 ms
64 bytes from xxx.76.32.153: icmp_seq=12 ttl=254 time=46.549 ms

-----------------------------------------
here is how ping to the 1st hop looks is when the link is empty

nat# ping xxx.76.32.153
PING xxx.76.32.153 (xxx.76.32.153): 56 data bytes
64 bytes from xxx.76.32.153: icmp_seq=0 ttl=254 time=16.468 ms
64 bytes from xxx.76.32.153: icmp_seq=1 ttl=254 time=7.100 ms
64 bytes from xxx.76.32.153: icmp_seq=2 ttl=254 time=7.491 ms
64 bytes from xxx.76.32.153: icmp_seq=3 ttl=254 time=6.313 ms
64 bytes from xxx.76.32.153: icmp_seq=4 ttl=254 time=8.211 ms
64 bytes from xxx.76.32.153: icmp_seq=5 ttl=254 time=8.101 ms
64 bytes from xxx.76.32.153: icmp_seq=6 ttl=254 time=7.499 ms
64 bytes from xxx.76.32.153: icmp_seq=7 ttl=254 time=39.151 ms
64 bytes from xxx.76.32.153: icmp_seq=8 ttl=254 time=5.856 ms
64 bytes from xxx.76.32.153: icmp_seq=9 ttl=254 time=6.396 ms
64 bytes from xxx.76.32.153: icmp_seq=10 ttl=254 time=7.066 ms

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

and here is an output also to the 1st hop ( running with similar altq
config ) during huge ftp transfer on 2mbps link

64 bytes from : icmp_seq=0 ttl=255 time=7.015 ms
64 bytes from : icmp_seq=1 ttl=255 time=6.845 ms
64 bytes from : icmp_seq=2 ttl=255 time=7.843 ms
64 bytes from : icmp_seq=3 ttl=255 time=7.964 ms
64 bytes from : icmp_seq=4 ttl=255 time=7.350 ms
64 bytes from : icmp_seq=5 ttl=255 time=6.362 ms
64 bytes from : icmp_seq=6 ttl=255 time=3.802 ms
64 bytes from : icmp_seq=7 ttl=255 time=4.915 ms
64 bytes from : icmp_seq=8 ttl=255 time=8.368 ms
64 bytes from : icmp_seq=9 ttl=255 time=5.353 ms

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

the same but no traffic on 2mbps

64 bytes from  icmp_seq=0 ttl=255 time=3.692 ms
64 bytes from  icmp_seq=1 ttl=255 time=1.962 ms
64 bytes from  icmp_seq=2 ttl=255 time=1.962 ms
64 bytes from  icmp_seq=3 ttl=255 time=1.938 ms

---------------------------------------------------------------------
heavy ftp without altqd on 2mbps
64 bytes from : icmp_seq=0 ttl=255 time=260.599 ms
64 bytes from : icmp_seq=1 ttl=255 time=6.151 ms
64 bytes from : icmp_seq=2 ttl=255 time=135.017 ms
64 bytes from : icmp_seq=3 ttl=255 time=7.581 ms
64 bytes from : icmp_seq=5 ttl=255 time=8.754 ms
64 bytes from : icmp_seq=6 ttl=255 time=141.742 ms
64 bytes from : icmp_seq=7 ttl=255 time=9.045 ms
64 bytes from : icmp_seq=9 ttl=255 time=123.588 ms
64 bytes from : icmp_seq=10 ttl=255 time=6.000 ms
----------------------------------------------------------
and here is config for this link - as you see very similar

interface fxp0 bandwidth 2.17M tbrsize 1000 priq
class priq fxp0 outgoing_udp_packets NULL priority 3
class priq fxp0 outgoing_ping NULL priority 2
class priq fxp0 ssh NULL priority 1
class priq fxp0 default_class NULL priority 0 default
filter fxp0 outgoing_udp_packets 0 0 0 27005 17
filter fxp0 outgoing_ping 0 0 0 0 1
filter fxp0 ssh 0 0 0 22 6
----------------------------------------------------------------------

As you can see , heavy ftp traffic makes no diffrence on icmp delay for this
2mps link but makes for this 128 kbps one.


Regards
References:
- [altq:1708] Re: PRIQ problem
  - From: Nicolas Christin <nicolas@cs.virginia.edu>
Prev by Date: [altq:1708] Re: PRIQ problem
Next by Date: [altq:1710] simple way to limit traffic with exception for dmz
Prev by thread: [altq:1708] Re: PRIQ problem
Next by thread: [altq:1702] Re: PRIQ problem (fwd)
Index(es):
- Date
- Thread