Subject: Re: NetBSD and PF+ALTQ revisited
To: Andy Ruhl <acruhl@gmail.com>
From: Mark Cullen <mark.r.cullen@gmail.com>
List: netbsd-users
Date: 01/08/2007 18:44:51
Andy Ruhl wrote:
> On 1/8/07, Mark Cullen <mark.r.cullen@gmail.com> wrote:
>> Hi,
>>
>> A while a go I was asking about how to get ALTQ working with PF, and one
>> way or another I ended up finding some patches to enable the
>> functionality. Said patches were from http://nedbsd.nl/~ppostma/pf/, but
>> this site no longer seems to have the 'beta' patches for NetBSD 3.x any
>> more. It just says to use -current now
>> (http://nedbsd.nl/~ppostma/pf/altq.html). This is something I don't
>> really want to do.
>
> Last I checked it's supposed to work with NetBSD 4. I guess we'll have
> to wait and see. I want to get it working too, but I don't want to use
> current or jump through hoops like you're doing.
>
> Andy
>
Well, this is sort of good news at least! I don't really want to jump
through hoops either, but it looks like I have to. I find it hard to
live without TCP ACK priority going :-) For what it's worth, I seem to
remember it being really easy to get the patches working with 3.0.1, but
that was admittedly a rather long time a go now!
I still hope it gets back ported to 3.x some time, if that's even possible!
Small update to the situation, the compile just finished, apparently
without error this time. Running `./pfctl -vs queue` from
/usr/src/dist/pf/sbin/pfctl seems to work fine, so I assume that, well,
it's going to work!
You might also wonder why I am bothering to update if 3.0.1 was working
OK for me? Well, yesterday, after ~90 days uptime I got some weird ATA
related messages. Something about viaide losing an interrupt, over and
over. Both my RAID-1 arrays had one disk marked as failed too. The
machine didn't panic, but wd2 turned totally inaccessible until restart
(atactl wd2 <any command> = timeout).
wd3 (on the same channel as wd2 from what I remember) is a Seagate 160GB
drive, and the change log for 3.1 does mention something about Seagate
driver related updates, so I am hoping this might be the solution to
this problem...
--
Mark Cullen <mark.r.cullen@gmail.com>
BSc (Hons), Computer Science