MC/SG Virtual IP Address stops responding

[Alfred J]

I have a monitored environment that connects to our ITO server using a MC/SG virtual IP address. The environment can ping this IP address only for a short time, from about 1 to 25 minutes, then the ping fails. We can get ping to work again by clearing the arp cache on the router closest to the monitored environment (vr1, see below), but again within a short time the ping fails.
The static IP addresses have no problems at all from the monitored environment.
There is no problem from vr2 or the 7204.
(One of our network folks mentioned something about a virtual mac address. I've never heard of this before. Can any one elaborate?)

The environment looks like:

monitored environment
|
+--|----------------------------+
|virtual router (vr1) |
| | cisco 6006 |
|virtual router (vr2) |
+--|----------------------------+
|
cisco 7204 router
|
ITO Management Server

[Terja]

mac address is the lans unique hardware address, however, since hp-ux 11 this has been possible to be software coded. a common trap is to do a restore, and blow away the previous value of the mac address. check lanscan, against what your network guys see as the mac address on the network.
also have you checked that there are no duplicat ip addresses, that your default router and mask are correct.

[Alfred J]

Network configuration is correct, and works as proved by it initially working, and after arp cache is cleared. There are no duplicate IP addresses. The 7204 router sees the correct mac address for the IP address(es). The other routers see the mac's of the next router interface in the path.

[Dieter Degrendele_1]

Hi,

We had the same problem when we changed the network cards on one node. We could ping the machine from some subnets, other subnets did not reach the node. After long time of discussions we decided to shutdown the switch and the problem was solved. We have the impression that something is wrong with ARP. Did you apply the latest firmware for your switches?

Rgds,
Dd

[Gengwd]

Hi Alfred,

Have you solved the problem? I want to share your experence, can you help me?

Regards

[Alfred J]

The problem was related partially to a misconfigured Sun Server which caused
an MLS entry (hardware based routing) for the HP/OV cluster address to be
overwritten with a incorrect entry. Once Nathalie fixed the incorrect route on the Sun
server, the switch was no longer getting confused.