[ale] SSH Tunneling strangeness

[Bob Toxen]

That is indeed strange.  I'll give a "wild" low-probability guess that
it might be the "TCP inside of TCP" problem with the different levels'
timeouts getting confused.  Would those commands that break it be
emitting large amounts of data in bursts?  Try disabling the ClientAlive.

Just a guess,
Bob

On Wed, Dec 17, 2003 at 12:22:43PM -0500, Matt Smith wrote:
> I've got several boxes in the "wild" that SSH into one central box, and do remote port forwarding to allow a back-door into themselves.  I.E. forwarding a given port on the central box to their local port 22.  I've got a while loop keeping the connection up, and using KeepAlive & ClientAlive settings in SSH when the connection drops when it fails and gets re-established pretty reliably.
> 
> Here's the problem.. on a couple of boxes, when I ssh BACK INTO the remote boxes (via the tunnel - ssh over ssh basically ), basically SSHing to localhost on a different port, the connection will die when I type certain commands.. It's doesn't appear to be the commands themselves that break it, but maybe the sudden data stream, or something in the return data that is breaking something.  Using debug on both ends I've found that basically the inital connection from the remote site to the central box stops responding to the ClientAlive checks from the central box and after the given timeouts, the connection is dropped from the central server.  The while loop goes and it reconnects..
> 
> I've searched and searched for any explanation for this, but have yet to find anything.  I've got a mix of various RedHat versions - from 7.3 to 9.  No obvious pattern for the ones that work fine vs. the ones that have this problem.
> 
> Any ideas??
> 
> --Matt
> 
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale