We recently experienced a problem with overnight jobs which were held up due
to a user-job unexpectedly requesting a tape mount after the operators had
gone home for the night. Because of this a backup was not done, and when the
prior backup DAT snapped during RESTORE, we found ourselves in an embarrassing
situation.
        This started me thinking about how we might have broken the chain of events
described above, and how nice it would be to have the HP3000 tell us if
certain unusual categories of events occurred. Ideally, in the above situation
it would detect the tape request not being serviced, and take some action -
for instance abort the job, or page an operator.
 
        Having recently installed the new SNMP capable OpenView /Windows 7.2 for our
DTCManager upgrade, I noticed that paging software was shipped with the
standard version of OpenView NodeManager, and that the network management
software could potentially be a centralized solution for reporting problems on
DTC's HP3000's, and HP9000's in our computing environment.
 
        On the HP3000 side, command procedures supplied with the SNMP component of
MPE/iX could be used to send a SNMP trap to the OpenView workstation. The only
missing piece is how to detect the abnormal events to be reported. Apparently
this must be possible, because HP's OpenView/Operations Centre product seems
to be able to do this, unfortunately, at over A$100,000 it is a few orders of
magnitude too expensive for our budget!
 
         The nearest I have come is an
Interex CSL program written in Turbo/Pascal which uses a PC to monitor console
messages and work a pager, so if anybody knows of a lights-out operations
monitor program for the 3000, resonably-priced which would help me avoid the
scenario described above, I would like to hear about it.
 
Thanks in advance,
 
Andrew McClaren
Systems Manager
RMIT, Melbourne - Australia
[log in to unmask]