Hi

I have a development area that utilizes rcp/rsh between the servers. Recently, we made a network change to have several VLANs instead of a Single-VLAN network to better simulate the actual productions environment.

After the change, all my servers are now connected via a Cisco 2960G switch. I can ping and do a traceroute to the servers but rcp connections are returning "No route to host" for two of the servers.

I have a total of 5 servers - A, B, C, X, Y. A and B is setup as a cluster, and only B is setup for rcp connection. C, X and Y have been setup for rcp as well. This is what I see:-

B can rcp to C and vice versa
B and C cannot rcp to X and Y
Y can rcp itself but cannot rcp elsewhere --> No route to host.
X can't even rcp itself --> It reports a usage error but the syntax is correct.

The closest source I have is below, but with IPtables turn on. My iptables are turn off for the setup.

RCP with IPtables

I also noted the following commands available on the Cisco switch:-
ip rcmd remote-host
ip rcmd rcp-enable

But from my reading, it only seems to allow rcp connection into the Cisco switch and not beyond.

Any suggestion would be appreciated.
I'm somewhat clueless about what's happening.

RCP uses files like hosts.equiv and .rhosts. Make sure you have them set up to allow rcp from the source servers

Thanks. I knew about this. Anyway, the initial problem have been resolved. There was an firewall up which I was not aware of and did not expect, so I had missed checking that out.

Nonetheless, I have an outstanding strange issue with the rcp command.

Server X can't even rcp itself --> It reports a usage error but the syntax is correct.

Here is the exact error:-

As mentioned, 192.168.2.8 is the node own address.
Note that under the error, it ALWAYS appear as "sage" with the "u" dropped. This same thing is seen when I tried to rcp from other servers to this one.

I re-installed the rsh and rsh-server package but I'm still getting this.

This server houses Veritas Netbackup. Could it be that NetBackup is screwing up the rcp connection?

Any ideas anyone?
Just to add that the server can rcp to other servers, but it can't received rcp connections and always throw the error above.
Thanks.

Edit: Resolved. Uninstalled and reinstalled the rcp a few times to no avail. Shut down NetBackup services, reinstall rcp and it works. Start up NetBackup services and rcp remain working. I don't quite understand how and why NetBackup would affect rcp. Probably some glitch in my setup. Anycase, everythings good now.