ssh -R

[Chris G]

On Tue, Apr 15, 2008 at 11:33:54AM +0000, tony.chamberlain at lemko.com wrote:
> 
> 
> 
>    The following is for CentOS 4.5
>    We have an internal network (192.168.5.0/255.255.255.0).
>    We have one machine reachable from inside and outside
>    (NOT on the 192.168.5 network).  Just for this example
>    call it 10.20.30.40 (though that is not its real address.
>    I don't put the real address, for security concerns here).
>    Anyway my machine is 192.168.5.19  so from my machine
>    I do an
>         ssh -l root -R 10022:127.0.0.1:22 10.20.30.40
>    Then I log into 10.20.30.40 from another machine and do a
>         ssh -l tony -p 10022 127.0.0.1
>    which gets me into my machine.  Test passes.  Problem is, by
>    the time I get home, my ssh -l root -R 10022:127.0.0.1:22 10.20.30.40
>    has timed out or something and I can no longer get to my local machine.
>    Do you know what I can do to keep it from timing out (or maybe locking up)?
>    I do have root access to both machines so if there is something in
>    sshd_config to change, I can do it.

I have a cron job which runs the script below every 15 minutes to see if
the ssh is still running and restart it if it isn't:-

    #
    #
    # Script to set up a secure tunnel from home system
    #
    cn=`ps -ef | grep "ssh -l chris -R 50022:apollo:22 -N xx.yy.zz.aa" |
    grep -v 'grep ssh'`

    if [ -n "$cn" ]
    then
        echo `date` "hssh is running" >/home/chris/tmp/hssh.log
    else
        /proj/chris/bin/ssh -l chris -R 50022:apollo:22 -N xx.yy.zz.aa
    fi

It means that even if there *is* a connection which has got screwed
up for some reason I can kill the ssh running on my home machine and
within 15 minutes the cron job and script above will start a new
session.

-- 
Chris Green