Unknown Behavior of PF+ALTQ on a Bridge

Rafael Henrique Faria rafaelhfaria at cenadigital.com.br
Thu Jun 24 23:01:38 UTC 2010


On Thu, Jun 24, 2010 at 18:56, Ermal Luçi <eri at freebsd.org> wrote:
> 2010/6/24 Rafael Henrique Faria <rafaelhfaria at cenadigital.com.br>:
>> Just to be more clean:
>>
>> My pf.conf:
>> ----
>> wan_if="bce0"
>>
>> set limit { states 100000, frags 20000 }
>> set loginterface $wan_if
>> set optimization normal
>> set block-policy drop
>> set fingerprints "/etc/pf.os"
>> set skip on lo
>>
>> altq on $wan_if cbq bandwidth 100% queue { out_bal, out_std }
>>   queue out_bal bandwidth 50% priority 0 cbq
>>   queue out_std bandwidth 50% priority 0 cbq (default borrow)
>>
>> pass out on $wan_if queue (out_bal)
>> ----
>>
> The problem is that this rule will not match any traffic that
> initiated as incoming on $wan_if.
>
> Try this instead:
>  pass out all queue (out_bal)
>
> It will do the magic.

I tried it... but nothing changes... the same behavior.

queue root_bce0 on bce0 bandwidth 1Gb priority 0 cbq( wrr root )
{out_bal, out_std}
  [ pkts:      76573  bytes:   14784373  dropped pkts:      0 bytes:      0 ]
  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
  [ measured:  2774.1 packets/s, 4.15Mb/s ]
queue  out_bal on bce0 bandwidth 500Mb priority 0
  [ pkts:      27413  bytes:    8197630  dropped pkts:      0 bytes:      0 ]
  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
  [ measured:  1040.4 packets/s, 2.34Mb/s ]
queue  out_std on bce0 bandwidth 500Mb priority 0 cbq( borrow default )
  [ pkts:      49160  bytes:    6586743  dropped pkts:      0 bytes:      0 ]
  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
  [ measured:  1733.7 packets/s, 1.81Mb/s ]

I have tried a lot of rules... including:

pass all queue out_bal

But without success...

If this is not the regular behavior of PF+ALTQ, my suspect is on the
Bridge itself...

>>
>> The "pfctl -vvs queue" show:
>>
>> ----
>> queue root_bce0 on bce0 bandwidth 1Gb priority 0 cbq( wrr root )
>> {out_bal, out_std}
>>  [ pkts:      50117  bytes:   13947411  dropped pkts:      0 bytes:      0 ]
>>  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
>>  [ measured:  3869.4 packets/s, 8.31Mb/s ]
>> queue  out_bal on bce0 bandwidth 500Mb priority 0
>>  [ pkts:      33198  bytes:    7175985  dropped pkts:      0 bytes:      0 ]
>>  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
>>  [ measured:  2591.3 packets/s, 4.36Mb/s ]
>> queue  out_std on bce0 bandwidth 500Mb priority 0 cbq( borrow default )
>>  [ pkts:      16919  bytes:    6771426  dropped pkts:      0 bytes:      0 ]
>>  [ qlength:   0/ 50  borrows:      0  suspends:      0 ]
>>  [ measured:  1278.1 packets/s, 3.95Mb/s ]
>> ----
>>
>> So, my question is: why the default queue is being used, If I have a
>> rule to use the out_bal queue to all outgoing traffic on that
>> interface?
>>
>> I need to redirect all the traffic from a subnet (/24) to one queue
>> (incoming and outgoing traffic)... so what I can understand is that,
>> this is not possible with PF+ALTQ. Am I wrong?
>>
>> --
>> Rafael Henrique da Silva Faria
>> Grupo de Sistemas e Redes
>>
>> Serviço Técnico de Informática
>> Faculdade de Ciências e Letras do Campus de Araraquara - UNESP
>> _______________________________________________
>> freebsd-net at freebsd.org mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-net
>> To unsubscribe, send any mail to "freebsd-net-unsubscribe at freebsd.org"
>>
>
>
>
> --
> Ermal
>



-- 
Rafael Henrique da Silva Faria
Grupo de Sistemas e Redes

Serviço Técnico de Informática
Faculdade de Ciências e Letras do Campus de Araraquara - UNESP


More information about the freebsd-net mailing list