networking on ucx 5.0A

[Matt Briggs]

HARDWARE: Compaq XP1000
OS: OpenVMS 7.1-2
UCX: 5.0A

Host was created by restoring a disk from another host, upgrading UCX to 5.0A, and changing the IP address to a free number.

We can ping all hosts on the local LAN, but get no response from the Cisco gateway address (and can't ping off the local segment). Other hosts on the LAN can not ping this host.

Ideas of what to check? THANKS!

[Andy Bustamante]

Some questions and speculation.

Verify the subnet and route,

$ TCPIP SHOW ROUTE
$ TCPIP show interface

just to make sure.

Is the gateway a firewall? or does it do any traffic filtering? If you can see the local LAN, and the subnet is correct the gateway may be ignoring your new system.

Do you have DECnet running on the this system? Is the system disk a copy of another host on the subnet? This will lead to duplicate MAC addresses if you didn't change the DECnet node address. All sorts of interesting behaviors could result.

And of course, do you have valid licensing for the copied system?


Andy

[Matt Briggs]

Thanks Andy. We are running decnet and ucx. We did create this host from restoring a disk from a different host. A correction to my previous - we can ping the new host from other hosts on the LAN, and it can ping them. Just can't get to (or through) the gateway, which is a Cisco firewall/router. One of the existing workstations can ping the gateway if we use its AUI interface, but can not ping using the UTP interface. On the new host, we can't get it to get a return from the gateway using the AUI or the UTP. When we remote into the gateway, it does show the IP address of the new host. It's like it know's the host is there, but will not respond to it, or the host does not hear the reply.... so at this point we're thinking it is more of a configuration issue on the gateway itself.

[Andy Bustamante]

Please execute

$ MCR NCP SHOW EXEC

on the new system and the original system. Are these systems on the same subnet? If the DECnet address (20.3 for example) is the same on both node this will be a problem. If the DECnet address is the same, that will force the MAC addresses to be the same.

Andy

[Jon Pinkley]

Matt,

If you are running DECnet IV, the MAC address is changed to correspond to the DECnet node number, as Andy Bustamante said.

If you put this node on the same "Ethernet" you will have problems. My guess is that isn't the case, or the other node that was cloned would also be affected.

From your descriptions, it seems highly likely that the Cisco is blocking, or ignoring you.

You don't specify a model, is this cisco a PIX, an ASA or a router with the FW feature set? Or is it some other Cisco device (Layer 3 switch?).

How did you determine the gateway knew about the ip address, from the output of the Cisco's ARP cache?

You also talk about AUI and UTP ports. Are these on the Cisco? Are they both connected to the same LAN segment? That sounds odd if they are.

We need a bit more info before we can make an informed guess as to what the problem is. If you can ping other devices on the local segment, and they can ping you, (with the exception of the cisco, which needs to be on the local network (as defined by netmask), then it is likely to be something in the cisco. On the other hand, if you have both Host using the same MAC address, and they are not seperated by a router, you will have problems no matter how the cisco is configured.

Jon

[M. T. Hollinger]

If you check the output from a TCPIP SHOW INTERFACE /FULL command, it will show you the Ethernet address for the interface in question. Does it start with AA-00-04?

Could you show us the output from the TCPIP SHOW ROUTE /DEFAULT command? Is the output different if you add the /PERMANENT qualifier?

Try to ping (or TELNET to) the default router. Does it work? Does the TCPIP SHOW ARP command show an Ethernet address for the router (if you issue the SHOW ARP right after a failed ping)?

Although it is probably not the answer to this particular problem, I would encourage you to upgrade to a more recent release of OpenVMS and TCP/IP Services.

- Mark

[Matt Briggs]

Thanks for everyone's help - we have it working now.

Here's more of the picture...we have a number of hosts that connect via AUI to a Delni. A Defar connects to the same Delni. The Delni connects to a Black Box hub; this hub is connected to a Cisco VLAN switch, that is in-turn connected to a Cisco router (the 'gateway').

The new host - that could not ping the gateway - has a UTP port that was connected through a UTP/AUI converter to the Delni. We removed the UTP/AUI converter, and connected the host directly to the Black Box hub via UTP. Now everything - annoyingly - works. So I still don't know for sure if the problem was caused by the Cisco gear declining to respond to calls from the host through the Delni, or if the UTP/AUI converter on the new host was the problem. As we looked into this more, we have also discovered that from outside of the LAN we can ping all devices on that LAN that communicate via AUI connector, and devices that are connected directly to the black box hub, but none that are on the far side of the Delni that use UTP. (my money is on the UTP/AUI converters at this point) Regardless, I appreciate everyone's assistance with the responses! THANKS!