Jose M. Prieto wrote:
> I have just found a good explanation of the problem [1]. It also explains the
> change made to "fix" the 2.6.8 kernel version.
>
> [1] http://lwn.net/Articles/92727/

Yeah, that's very helpful. Thanks.

> > I would suggest people to use this trick instead of purely disabling Window
> > Scaling:
> >
> >     echo 4096 87380 174760 > /proc/sys/net/ipv4/tcp_rmem
> 
> This does not completely disable window scaling, but lowers the default scale 
> factor.

Indeed it's only a workaround and it's not bulet-proof.

As far as I'm concerned, I now completly disable windows scaling because this
morning I experienced some issues. Moreover, bye only changing the scale
factor, you do *not* ensure that you can actually communicate with every host.
There might still be some unreachable hosts, even if you're not aware of them.

> So I get much better network performance disabling scaling than with the
> tcp_rmem setting.

Not for me as far as I know.

> I think disabling window scaling is the best option. Anyways, if others agree 
> that the tcp_rmem setting is better, I have no objection. My main point was 
> that Arch should provide a workaround by default.

For sure. But if it's done, we should let the user know about it. I'm affraid
of that Arch include a lot of workaround I'm not even aware of.

I'm an ex-Slacker and that's not something I appreciate much.

NB: It seems that your mail never arrived on the mailing list. It was marked as
SPAM on my box too.

-- 
Benoit Myard <[EMAIL PROTECTED]> [0x3EEC1AEC]

Attachment: pgpU6GB8JTflu.pgp
Description: PGP signature

_______________________________________________
arch mailing list
[email protected]
http://www.archlinux.org/mailman/listinfo/arch

Reply via email to