Svavar Örn Eysteinsson wrote:
> I was actually thinking about disable'ing NetworkManager and the avahi- 
> daemon.
> But wasn't sure about if I need it it or what.
> 
> BUT, I got some great great news.
> 
> I told my friend yesterday to start a download from my FTP server.  
> Everything went smooth.
> No crash, no nothing. So I decided to try it out my self just to make  
> sure everything was OK.
> Connected my laptop to my public network switch and gave it a legal  
> public IP from my network.
> Started a download from the ftp server and got about 10.8 - 10.9 MB/ 
> sec in download rate.
> Did that for about 20-30min or so, and no logs or no crash on my  
> firewall or router machine.
> 
> I also downloaded while my friend was downloading too, more stress on  
> the link right ?
> 
> So, I would think that the problem is gone.
> With updated drivers (8.0.16) and configuration ( the  
> TxDescriptorStep ) on both firewall and my router machine, thanks to  
> you David.
> And replaced a 100Mbit switch to a 1GB switch ( unmanaged although : 
> (    )
> 
> I'm still going to try more of a stress test on my link after the  
> weekend.. This situation is too good to be true :)
Svavar, this is great news, as you say maybe too good.
> Here's a simple and ugly diagram of the network in PDF format :
> http://dl.getdropbox.com/u/16134/g37-network-diagram.pdf
Thanks, I've had a look. Do I understand correctly, that the FTP get 
that seemed to cause the original problem ddidn't run its traffic 
through the AMD firewall that reported the TX Hang, but from FTPSrv to a 
public internet client ?

------ PUBLIC -------------------|-FireWall-|---PRIVATE---

             Zebra                    AMD
Internet  0+-----+ 1    +----+   0 +-----+ 1   +
& ftp    --|Linux|------| SW |-----|Linux|-----+ Lan
client     +-----+      +-+--+     +-----+     +
                           |
                        +--+---+
                        |FTPSrv|
                        +------+

If so, the FTP transfer must have effected the AMD firewall box 
indirectly, and that might be because the (10/100) switch was 
overloaded, configured for flow control, and sent flow control frames 
(PAUSE) out to all connected interfaces including eth0 on the AMD 
firewall. And if this was the case, then replacing the 10/100 switch 
with a 1G switch may really be what resolved the issue.

Let me know if I do correctly understand the network configuration.

Because the original problem you reported has gone, I don't feel that 
this is now as urgent as it was, but I'd really like to understand which 
  things did what.

Thanks
Dave

> 
> 
> 
> 


------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay 
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel

Reply via email to