headers
Alexander Leidinger | 1 Dec 2002 14:49
Favicon

Re: traffic prioritization.

On Fri, 29 Nov 2002 17:05:11 -0500
"Peter Brezny" <peter <at> skyrunner.net> wrote:

> I feel sure that it can be done with it.  I'm vuagely familiar with the
> etinc.com's bwmgr package which seems to do traffic prioritization.

Define a pipe and some queues which feed their data into the same pipe,
the data of the queues is then priorized depending on the weight of the
queues.  The data with the lowest priority doesn't get stopped, it will
find it's way through the pipe, it's just that higher priorized data
gets more of the max. bandwith of the pipe than lower priorized data.

Bye,
Alexander.

--

-- 
            Secret hacker rule #11: hackers read manuals.

http://www.Leidinger.net                       Alexander  <at>  Leidinger.net
  GPG fingerprint = C518 BC70 E67F 143F BE91  3365 79E2 9C60 B006 3FE7

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message
Permalink | Reply | 2 comments |
headers
Marcin M. Jessa | 2 Dec 2002 10:15
Favicon

Re: bandwidth management package.

Try DUMMYNET. It's a part of ipfw and it's avaliable "out of the box" on
FreeBSD. Both ezunix.org and bsdvault.net have articles about how to make it
work. 

Cheers.

YazzY

On (29/11/02 17:27), Peter Brezny wrote:
> From: "Peter Brezny" <peter <at> skyrunner.net>
> To: <freebsd-isp <at> freebsd.org>
> Subject: bandwidth management package.
> Date: Fri, 29 Nov 2002 17:27:00 -0500
> 
> What are you guys using for bandwidth management.
> 
> I've been looking at the etinc.com bwmgr package, but that's really all i've
> looked at.  Are there other bsd based competitors out there, better or
> worse, or different?
> 
> is it a huge deal to roll your own with ipfw2?
> 
> TIA
> 
> 
> Peter Brezny
> Skyrunner.net
> 
> 
>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Peter Brezny | 2 Dec 2002 18:07
Favicon

RE: traffic prioritization.

Alexander,

Thanks a bunch for the tip.

Looks like I need to do some more reading in the man pages!

Do you have an example of what you described below that I could look at to
minimize syntax battles?

Thanks again,

Peter Brezny
Skyrunner.net

-----Original Message-----
From: Alexander Leidinger [mailto:Alexander <at> Leidinger.net]
Sent: Sunday, December 01, 2002 8:50 AM
To: Peter Brezny
Cc: freebsd-isp <at> FreeBSD.ORG
Subject: Re: traffic prioritization.

On Fri, 29 Nov 2002 17:05:11 -0500
"Peter Brezny" <peter <at> skyrunner.net> wrote:

> I feel sure that it can be done with it.  I'm vuagely familiar with the
> etinc.com's bwmgr package which seems to do traffic prioritization.

Define a pipe and some queues which feed their data into the same pipe,
the data of the queues is then priorized depending on the weight of the
queues.  The data with the lowest priority doesn't get stopped, it will
(Continue reading)

Permalink | Reply | 1 comment |
headers
Alexander Leidinger | 2 Dec 2002 18:30
Favicon

Re: traffic prioritization.

On Mon, 2 Dec 2002 12:07:30 -0500
"Peter Brezny" <peter <at> skyrunner.net> wrote:

> Do you have an example of what you described below that I could look at to
> minimize syntax battles?

---snip---
############
# Flush out the list before we begin.
${fwcmd} -f flush
${fwcmd} -f queue flush
${fwcmd} -f pipe flush

${fwcmd} pipe 1 config bw 64kbit/s queue 10KByte
${fwcmd} queue 1 config pipe 1 weight 10 queue 10KByte
${fwcmd} queue 2 config pipe 1 weight 1 queue 10Kbyte

${fwcmd} add queue 1 tcp from ${inside_net}/${inside_mask} to any via ${interface} out
${fwcmd} add queue 2 udp from ${inside_net}/${inside_mask} to any via ${interface} out
---snip---

This adds a pipe with a limit of 8k. Both queues share this bandwith, but
one gets more than the other.

Then outgoing tcp traffic flows through queue 1 and udp traffic through
queue 2.

Bye,
Alexander.
(Continue reading)

Permalink | Reply | 0 comments |
headers
itchibahn | 3 Dec 2002 04:46

Re: DNS changes not take effect

Yes, serial is increased on each modification.
Turns out, it wasn't resolving forward at all on zones included in the 
named.conf.  If I remove all zones from the named.conf, it resolves fine, 
probably getting data from other DNS servers.

Quoting Ralph Forsythe <rf-list <at> centerone.com>:

> You increasing the serial on the zone file?
> 
> At 02:13 AM 11/29/2002 -0600, itchibahn wrote:
> >I'm using FreeBSD 4.6.2 and BIND 9.2.1.
> >
> >Whenever I add a record, the reverse resolves fine, but the forward does 
> >not on
> >the host.  It's like the cache is not being refreshed, after restarting
> >the 'named' or restarting the entire server.
> >
> >But then, why does it resolve fine on reverse but not forward?
> >
> >-------------------------------------------------
> >This mail sent through ISOT.  To find out more
> >about ISOT, visit http://isot.com
> >
> >To Unsubscribe: send mail to majordomo <at> FreeBSD.org
> >with "unsubscribe freebsd-isp" in the body of the message
> 
> 
> To Unsubscribe: send mail to majordomo <at> FreeBSD.org
> with "unsubscribe freebsd-isp" in the body of the message
>
(Continue reading)

Permalink | Reply | 1 comment |
headers
Alexandre Kardanev | 3 Dec 2002 08:15
Picon
Picon

Re: DNS changes not take effect


It can be in case of some errors in your zone files. Did you check
messages from named in syslog?

On Mon, 2 Dec 2002, itchibahn wrote:

> Yes, serial is increased on each modification.
> Turns out, it wasn't resolving forward at all on zones included in the 
> named.conf.  If I remove all zones from the named.conf, it resolves fine, 
> probably getting data from other DNS servers.
> 
> 
> Quoting Ralph Forsythe <rf-list <at> centerone.com>:
> 
> > You increasing the serial on the zone file?
> > 
> > At 02:13 AM 11/29/2002 -0600, itchibahn wrote:
> > >I'm using FreeBSD 4.6.2 and BIND 9.2.1.
> > >
> > >Whenever I add a record, the reverse resolves fine, but the forward does 
> > >not on
> > >the host.  It's like the cache is not being refreshed, after restarting
> > >the 'named' or restarting the entire server.
> > >
> > >But then, why does it resolve fine on reverse but not forward?
> > >
> > >-------------------------------------------------
> > >This mail sent through ISOT.  To find out more
> > >about ISOT, visit http://isot.com
> > >
(Continue reading)

Permalink | Reply | 0 comments |
headers
Troy Settle | 3 Dec 2002 12:34
Favicon

RE: dial-in recommendations please


Eric,

If you're not concerned with v92, you can pick up a Max 6096 on ebay for
less than $2000, or a 4048 for less than $1000.  You can also get
portmasters for less than $1000.

If you expect to have more than a few hundred ports, you may want to
look at the Max TNT, which can be had for a few grand + ~$1500 per 96
port modem card.

If you want something new, check out www.patton.com.  Everything I've
heard about these guys is good stuff.

--
  Troy Settle
  Pulaski Networks
  540.994.4254 - 866.477.5638
  http://www.psknet.com

> -----Original Message-----
> From: owner-freebsd-isp <at> FreeBSD.ORG 
> [mailto:owner-freebsd-isp <at> FreeBSD.ORG] On Behalf Of Eric 
> Brunner-Williams in Portland Maine
> Sent: Friday, November 22, 2002 10:47 PM
> To: freebsd-net <at> FreeBSD.ORG
> Cc: freebsd-isp <at> FreeBSD.ORG
> Subject: dial-in recommendations please
> 
>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Ralph Forsythe | 4 Dec 2002 02:10

RE: dial-in recommendations please

I just picked up a USR TC chassis fully populated with v.90 flash for
$180 from ebay, for reference.

On Tue, 3 Dec 2002, Troy Settle wrote:

>
> Eric,
>
> If you're not concerned with v92, you can pick up a Max 6096 on ebay for
> less than $2000, or a 4048 for less than $1000.  You can also get
> portmasters for less than $1000.
>
> If you expect to have more than a few hundred ports, you may want to
> look at the Max TNT, which can be had for a few grand + ~$1500 per 96
> port modem card.
>
> If you want something new, check out www.patton.com.  Everything I've
> heard about these guys is good stuff.
>
> --
>   Troy Settle
>   Pulaski Networks
>   540.994.4254 - 866.477.5638
>   http://www.psknet.com
>
>
> > -----Original Message-----
> > From: owner-freebsd-isp <at> FreeBSD.ORG
> > [mailto:owner-freebsd-isp <at> FreeBSD.ORG] On Behalf Of Eric
> > Brunner-Williams in Portland Maine
(Continue reading)

Permalink | Reply | 0 comments |
headers
Hernan Nunez | 4 Dec 2002 18:16
Picon

Sendmail + Milter + Amavis-Milter

Hi,

 I would like to setup Sendmail + Milter-ng + Amavis with milter interface.

I've been build my sendmail with the -DMILTER option then when i try to
compile amavis + milter i get an error with the shared library pthread.. Any
idea ???

Thanks in advance ...

cc  -DAMAVISD_SOCKET=\"/var/run/amavis/milter.amavis\"  -DRUNTIME_DIR=\"/var
/spool/amavis\"  -DPID_FILE=\"/var/run/amavis/amavis-milter.pid\"  -o
amavis-milter amavis-milter.c  -L/usr/lib/libmilter/ -lmilter -lpthread
amavis-milter.c:129: warning: static declaration for `strlcpy' follows
non-static
/usr/libexec/elf/ld: cannot find -lpthread
*** Error code 1

#################
# pthread library path
#################
ls -la /compat/linux/lib/pthread

###############################
# Linux Compat is Enabled and Working
###############################
kldstat
Id Refs Address    Size     Name
 1    4 0xc0100000 399530   kernel
 2    1 0xc2700000 18000    ipl.ko
(Continue reading)

Permalink | Reply | 5 comments |
headers
Peter Brezny | 4 Dec 2002 20:29
Favicon

avoiding circular route problems when changing data providers (advice needed)

Hi everyone.

I need some way to set up preferred routes based on source addresses.

We're migrating from one data provider to another, and I'd very much like to
simply put the new address range on the same freebsd router I'm using now
and plug it's primary interface into a switch that's connected currently to
both data providers routers (the old and the new).

Doing so leaves data from the new provider's ip space on our network heading
out to the internet via the old data provider's router due to the existing
default route.  And then coming back through the new provider's router.  I'm
not actually sure this is going to be a huge problem, unless the old data
provider won't allow traffic to pass their network unless it's from their
address space, which I have yet to test.

Is there a way that I can specify if the source address of the packet came
from the new data provider's ip block, go out the new data provider's
router?

Could it actually be as simple as:

ipfw add fwd all from x.x.x.x/23 to x.x.y.1

where x.x.x.x/23 is the new network range and x.x.y.1 is the new router.

Could this work?

Just for more info, we've got multiple cards in the freebsd router all with
traffic headed out through a single card.
(Continue reading)

Permalink | Reply | 2 comments |

Gmane