tinc hangs on large quantities of data

[Guus Sliepen]

On Thu, Jan 10, 2002 at 03:02:29PM -0600, ebina1 wrote:

> I have tinc setup between three machines debian linux 2.2.17, redhat linux
> 2.2.14-5.0smp, and redhat linux 2.4.2-2.
> 
> Everything works unless I try to transfer large amount of data from
> the 2.4.2-2 machine to either of the others.  When I do that
> connection hangs.  Something as simple as telneting to that machine
> from on of the others and typing 'ls -l' in a large directory like
> /bin will hang the telnet session.  Likewise with ftp it hangs the ftp
> session.  The other direction or between the other two machines works
> fine.  
> 
> Seems like an OS bug.  Anyone else using tinc with this version of redhat
> and having problems?

One of the problems that can cause this is a firewall that drops
fragmented UDP packets. The TCP layers will try to resend those packets,
without ever knowing whether the packets were dropped because of this or
that the packets were just lost on the internet.

You can try to solve this problem by decreasing the MTU of the ethertap
or tun/tap interfaces. You can do this by changing the tinc-up script
and append "mtu 1400" to one of the ifconfig commands. Whether or not
this fixes your problem, please let me know the results!

-- 
Met vriendelijke groet / with kind regards,
  Guus Sliepen <guus at sliepen.warande.net>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: not available
Url : http://brouwer.uvt.nl/pipermail/tinc/attachments/20020110/b51572a1/attachment.pgp