A remote host refused an attempted connect operation.

is their a special reason why you want to use rsh/rcp?
I'd suggest that you use ssh/scp instead.

from aix to linux you're getting the rsh/rlogin because r* services are defaulted to off on most linux distros for being insecure (cleartext, just like telnet). ahoffman's suggestion to use ssh/scp instead of rsh/rcp is an excellent idea. See http://unix.ittoolbox.com/documents/popular-q-and-a/installing-ssh-on-aix-2475 for how to install openssh on AIX if it is not alreayd installed (first link there is http://unix.ittoolbox.com/groups/technical-functional/ibm-aix-l/212785 which will be what you're looking for I think).

Timely topic.
I am experiencing the same issue migrating from Aix 4.3.3 to Fedora Core 4.
Also at another client I will be migrating from AIX 5.1 to Enterprise Linux AS 4.
At a third client we will be migrating from DG (unknown OS version) to AIX 5.3.

The problem is some of these older machines do not have ssh / ssl installed and in some cases are so ancient that it is difficult to configure ssh and ssl on them. The point is to migrate off them as quickly as possible. Therefore, running scp or ssh from the older machine to the new Linux boxes is not an option. You could configure the new machine to allow these commands to work; my solution is to pull the data from the old machine to the new one. In that manner, you do not compromise security on the new machine.

Which brings us back to the original question.

Bottom line is the command was successful and accomplished the copy. It was just squawking about your lack of attention to security issues. I have tried several things to eliminate these messages and have not been successful.

I have observed several such commands in operation recently and they seem to attempt connections from the most secure to the least and report failures as they descend.

Not to worry, the file was copied.

If you are planning to move entire directory structures, it is possible to migrate them without making an interium copy. The trick is to create a tar image on the source machine and pipe it to a tar command on the target machine.

The following worked on my configuration. The pontiac server is running Linux Fedora Core 4 and the chevrolet server is running AIX 4.3.3.

[root@pontiac junk]# pwd
/brc/junk
[root@pontiac junk]# rsh chevrolet '(cd /brc/sg; tar cf - try4)'|tar xvf -
Trying krb4 rsh...
krb_sendauth failed: You have no tickets cached
trying normal rsh (/usr/bin/rsh)
try4/
try4/f4.3
try4/f4.4
try4/f4.1.Z
try4/f4.2.Z
try4/f4.5
[root@pontiac junk]# ls -lR
.:
total 4
drwxr-sr-x  2 root sys 4096 Mar  9 06:44 try4

./try4:
total 1072
-rw-r--r--  1 root sys 1085668 Mar  9 06:34 f4.1.Z
-rw-r--r--  1 root sys    2247 Mar  9 06:36 f4.2.Z
-rw-r--r--  1 root sys       0 Mar  9 06:40 f4.3
-rw-r--r--  1 root sys       0 Mar  9 06:41 f4.4
-rw-r--r--  1 root sys       0 Mar  9 06:44 f4.5
[root@pontiac junk]#

I would change the xvf to xf on the rsh command to minimize screen output.

The above will work fine. However, I have experienced some issues you should watch for.

1)      The machine must both be configured to process files larger than 2GB if that applies.
2)      If there is a network switch involved, the communication link can cause collisions if the network interface is set to auto negotiate. I have had to set them to 100 or 1000 and turn off auto negotiate on both the switch and the AIX interface. The more data you are pumping, a bigger the pipe will make a difference.

Good Luck