HP OpenVMS Systems

Ask the Wizard Questions

Satellite boot problem

The Question is:

We have a cluster with 3 VAX and 5 Alpha machines with one
node each serving as the boot node for it's corresponding
architecture.  Strangly, after the most recent shutdown,
the Alpha boot node no longer responds to ethernet boot
requests from it's Alpha satellites.  All VAX nodes and the
single Alpha boot node are booted and in the cluster.
I can see the ethernet boot requests on the VAX boot node
with the following message.

%%%%%%%%%%%  OPCOM  21-DEC-1995 11:15:18.99  %%%%%%%%%%%
             (from node UCDHED at 21-DEC-1995 11:15:17.94)
Message from user DECNET on 
DECnet event 0.7, aborted service request
From node 41.422 (), 21-DEC-1995 11:15:17.94
Circuit SVA-0, Line open error, Ethernet address = 08-00-2B-39-B9-16
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

The VAX decnet database only contains load parameters for
the VAX nodes so this is what I expect.  The Alpha boot
node is absolutely silent and seems not to recognize the
boot requests at all.

As a test of the ethernet connection, I add the Alpha decnet
load parameters to the VAX boot node database (which is
responding to the boot request with the above error message).
so that LOAD FILE, and LOAD ASSIST point to the correct
system disk on the unresponsive Alpha boot node. Lo and
behold, the Alpha satellites boot using the VAX boot node
to point to the remote Alpha disk, so the ethernet
connection must be working.  I'm baffled and would be very
interested to discover the source of this behavior.

       Thanks,

The Answer is:

This really sounds like the Alpha boot node has had MOP service turned off
on it's ethernet circuit.  Execute CLUSTER_CONFIG on this boot node and
make the system a boot server.  Also make sure this system still has the
boot information for it's satellites.