Ask the Wizard Questions
Update: Wrong ACP
The Question is:
I have been trying to upgrade an antiquated VMS 5.2
machine (a Vaxstation 3500) to run VMS 5.5. As we have a
site license for VMS I tried what I thought was a simple solution
and copied the system from a machine already running VMS 5.5.
Things seem to be OK after I fixed up the node specific
stuff but I now get the queue manager hanging when it is started
or when a job is submitted to it.
The operator log messages are as follows:
%QMAN-E-COMMERROR, unexpected error #5 in communicating with node CSID 00000000
-SYSTEN-F-WRONGACP, wrong ACP for device
The first message is not listed in my system messages
document and I haven't got the faintest where the CSID 00000000
comes from. DECNET starts fine so the networking host info
is mostly correct.
What is QMAN trying to do?
The Answer is:
Assuming you do not want to (or need to) preserve the existing
queue definitions -- that the startup files are set up to create
the queue definitions if necessary, simply reinitialize the queue
manager database and recreate the necessary batch, print and server
queues. See START/QUEU/MAN/NEW.
The queue manager uses SCS information to keep track of queues
across a VMScluster, and it would appear the queue database here
contains "stale" SCS information, information associated with the
My tendency here is to use the /ON qualifier on the INIT/QUEUE/START
commands, and always passing in the current SCS nodename value -- the
value retrieved via a f$getsyi("SCSNODE") specification. This ensures
that the correct (current) node information is always specified. Each
node -- standalone or present in a VMScluster -- issues the queue
commands associated with queues local to itself, using this syntax.