Solved: xinetd refuse connect

[bxf4@psu.edu (Brian P. Flaherty)]

I am cutting out the past writing in the interest space.  The answer
turned on the xinetd entry, hosts.allow, and hosts.deny.  Thanks
Mike, for suggesting xinetd configuration issues.  I was thinking in
terms of one machine providing the proper information to another.

I had the wait line in the rsync entry set to yes, and that was
related to the <no address> message I was getting in the logs.  When I
changed that entry to no, rsync started getting my laptop's ip.  I
thought limiting wait to yes would limit the number of connections.

Also, I had called the service rsyncd, but it must be rsync and then
rsync must be the daemon name in hosts.allow.  Finally, the rsync call
in the server_args field of xinetd.conf must look like this "rsync
rsyncd --daemon".  I don't understand why this is, but it is.

Here is my working xinetd.conf entry for rsync.

service rsync
{
         flags        = REUSE NAMEINARGS
         socket_type  = stream
         protocol     = tcp
         wait         = no
         user         = root
         port         = 873
         server       = /usr/sbin/tcpd
         server_args  = /usr/bin/rsync rsyncd --daemon
         only_from    = 127.0.0.1 192.168.2.25
         log_on_success += DURATION HOST USERID
}

The only_from above is redundant with hosts.allow.  Here is the
relevant portion of my hosts allow:

rsync: 127.0.0.1 192.168.2.25

Have a good night.

Brian Flaherty