Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 9 Jul 2003 14:00:25 -0700 (PDT)
From:      Julian Elischer <julian@elischer.org>
To:        Kenneth Culver <culverk@yumyumyum.org>
Cc:        John Polstra <jdp@polstra.com>
Subject:   Re: wierd dsl performance with -CURRENT
Message-ID:  <Pine.BSF.4.21.0307091353440.22588-100000@InterJet.elischer.org>
In-Reply-To: <20030709164151.A5579@alpha.yumyumyum.org>

next in thread | previous in thread | raw e-mail | index | archive | help


On Wed, 9 Jul 2003, Kenneth Culver wrote:

> > Can you do a
> > 'netstat -s -p tcp >> tcpstats' before and after the transfer?
> >
> > This should tell us if there were retransmits etc.  It could be a
> > difference in minimum rtt values or a congestion issue that results in
> > timeout for our stack but some other recovery mechanism with other stacks.
> >
> 
> Here is the output of the netstat before and after transferring "kern.flp"
> for FreeBSD 5.1:
> 
> tcp:
[...]
>         35600 packets received

[...]
>                 1220 out-of-order packets (1657153 bytes)
> 
> This is after:
> 
> tcp:
[...]
>         36688 packets received
>                 1298 out-of-order packets (1770089 bytes)

This is the only thing that is of any interest..
and it could only really be if the etherent driver was re-ordering
them..

possibly FreeBSD 5  might react more to out-of-order packets..

capture a download with tcpdump 
and save it (ascii version) to a file using the -ttt option to get
relative timestamps..

look of any large values in the timestamps and see if there is anything
before that indicates a lost packet or a re-ordered one or something
(or a retransmitted ack)

The key is to find the gap in the arriving packets and figure out 
what caused it..

> 
> Thanks...
> 
> Could this be something my ISP did that causes a problem for FreeBSD and
> not the other OS's?
> 
> Ken
> 



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0307091353440.22588-100000>