UCX: 4.2 - cannot reach itself

[Willem Grooters]

On a VMS 7.1-2 + UCX 4.2 machine, UCX PING, TELNET or FTP to itself fails - acutally, time out.
Remote access however is no problem at all, nor is access to other machines (ping, telnet or FTP). It's just the access to it's own interface that is impossible.
Services TELNET and FTP are standard and enabled, and ROUTING is enabled and running. The network interface is Ok as well.

[Willem Grooters]

A sudden idea crossed my mind, I've seen something likewise before: it could have something to do with routing.
UCX SHOW ROUTE and yes: in the dynamic table I found address 10.100.50.201 routed via 10.100.50.1. Indeed, 10.100.50.201 is the machine itself, 10.100.50.1 is the router.

UCX SET NOROUTE 10.100.50.201 /gateway=10.100.50.1 removed the line, and to be sure we did the same /PERM.
Then it worked again, as expected.

Ok, problem solved (quite a pity I can't assign points to myself ;-)) but still a question to be answered: Since the entry occurred in the dynamic table, it's cause is not the permanent database, so where did it come from?

[Martin P.J. Zinser]

Obviously this is all gueswork, but would it be possible that another system was in the network using the same IP address? That way the router could have learnt and advertised it.

[Willem Grooters]

Missing bits & pieces in some DHCP server? Given the environment not unlikely ;-). It might shine a light on other problems I've heard talkink about. I'll inform them...Thanx.

[Stanley F Quayle]

By "itself", are you meaning the IP address you've assigned it?

Is that name and address in the local database?

$ UCX SHOW HOST xyz

You might also try various IP operations to address 172.0.0.1, which always means the local machine.

I agree that a different machine might also have the same IP address.

[Martin P.J. Zinser]

I think Stan is refering to 127.0.0.1 not
172.0.0.1...

[Stanley F Quayle]

Sorry. Brain fade...

[Willem Grooters]

Accessing 127.0.0.1 (localhost) was not a problem, it really was when the Ip-address of the machine in the network itself (10.100.50.201) was accessed - by address or by name (DNS was ok ;-))
Anyway, routing shouldn't have been enabled anyway so it'll be disabled. For the rest: all is fine now.

(BTW: Today I learned the machine had been up and running for over 3 years - doing nothing in fact. It was only that they discovered it couldn't 'see' his own interface that was decided to reboot - which didn't solve the problem ...)