Conference csc32::consolemanager

Hi,
   part of PCM is a 'C' language API and as part of this there is a call which
allows a program to send events into the ENS part of PCM. As a test for this API
I wrote a small program which takes events from the PSWX (ULTRIX Watchdog) and
fed them into ENS using the API calls. This is NOT a supported part of the code
as it was an exercise for me ( in C programing ) and used as a check for the API
documentation. Hopefully I still have this code and if you want it to play with
you are welcome. Let me know.

Cheers SImon...

    Hi,
    
    A complete PSW to PCM Feeder will be provided (and supported) in the
    next release of OSCint which is the 2.0 one and should follow as close 
    as possible PCM official release.
    
    Cheers,
    Christian.

    
Thanks Simon, but my customer uses OpenVMS/(VAX/AXP) systems, so the program you
wrote for Unix system is not usefull for me (unless the global design and
idea).

I am more interested in the new release of PCM and OSCint of which Christian 
told in his reply.

I catch the occasion to propose a problem arised by my customer.

The customer is still using VCS and PSW, but I think he will move to PCM and
PSW in the next time.

Actually PSW and VCS communicate using a mailbox and the C program is available 
in the VCS$EXAMPLES directory.

The message, sent via mbx by the PSW to ENS, has (more or less) the following 
information:

Node name, who generated the event
date & time
Event text
priority
Event name
Information

This is the customer example to explain better which are his needs.

The batch queue has been stopped. So the PSW sent a message to the ENS:

Node:        VAX1
Date & time: 10-jan-1994 12:04:00
Event name:  sns_c_bat
Information: sns_event
...

Then the PSW sent a new message to the ENS:

Node:        VAX1
Date & time: xx-xxx-xxxx xx:xx:xx
Event name:  sns_c_bat
Information: sns_event
...

To verify if the message is a new trouble message or if it is a 
"start recovery procedure" message the customer must look at the "date & time"
field.

If the value of the field "date & time" are the same in both messages then the 
messages are concerning the same trouble and the former is the trouble 
notification, the latter is the notification of the start of the recovery 
action.

If the value of the field "date & time" are different then the two messages 
are concerning two different problems.

What my customer is asking for is a solution to change the contents of the
field "information" when the action routine has started. He suggests something
like "sns_restart".

Maybe the C program can be modified, but the customer is scared of changing
something that is working. And moreover the C program has not many comments
inside.

Can you give me any suggestion or give me the good new "New version of PCM can
do that!"?

Thanks in advance.

							Anna
                                                                             

    Re -1.
    
    Very minor nit but the code VCS$FEED_ENS.C does not exist, rather it
    is SNS$FEED_ENS.C and it is part of System Watchdog. I say this only to
    avoid confusion for others that are reading this notes string.
    
    >Can you give me any suggestion or give me the good new "New version of
    >PCM can do that!"?
    
    There are two solutions here as I see it:
    
    1. You can modify the "feeder code" to have the smarts to send
       different strings to signal that the action routine has started.
    
    2. System Watchdog should be modified to accomodate the situation.
       I feel this solution is best as not all sites that use System
       Watchdog use VCS or will use PCM.
    
    Regs,
      Dan