rcp through Raptor Firewall over 2 domains

[Chris Treharne]

When trying to rcp any file from one server to another it works once and then fails with a message of "crmd: lost connection". When I leave the server for a minute or so and try again the rcp works once, and the failure cycle is repeated. rpc works fine between this machine and others on the same subnet.
rcp has been enabled on the firewall and telnet, ftp, rlogin etc all work fine with no problem.

Does anybody have any ideas please???

[Berlene Herren]

Is this an 11.0 system Chris? If so, check and make sure PHNE_20436 and/or PHNE_20735 are NOT installed on the machine. They introduced a new algorithm which results in a
much more random Initial Send Sequence number in TCP SYN and TCP SYN/ACK packets. From a security standpoint, this is very desirable, unfortunately, for applications that connect repeatedly to the same service on the same remote system, this can lead to periodic delays or failed connection attempts.
- Commands like remsh(1), rlogin(1) and rcp(1) can suffer from this. The only solution is to restart the operation
or to remove the patches from the system.

PHNE_21767 should be installed after these patches are removed.


Berlene

[Chris Treharne]

Thanks for the reply. I tried what you suggested but still no joy. Any further suggestions would be more than welcome. Thanks.

[Berlene Herren]

some things to check:

your /etc/passwd file >> make sure spellings are correct

your /var/adm/inetd.sec >> is there a deny statement in there? Try moving it to /var/adm/inetd.old and rcp again.

Cbeck /etc/inetd.conf for the login. It should look like this:
login stream tcp nowait root /usr/lbin/rlogind rlogind

If you have to make any changes, ensure you re-read it with inetd -c.

If you have a multihomed host, run inetd -l on host B and rcp from host A to host B, and observe the var/adm/syslog/syslog.log for any connections from "unknown." There maybe a problem with the netconf file.

Berlene