Conference csc32::consolemanager

    Tony,
    
        Did you assign a scan profile for each system in the configuration 
    editor? 
    
    Dave

This PCM system has been mentioned several times in this conference before by
Steve McGinley. We have 83 managed systems, and have been working well for many
months. I have just started trying to manage it, and it went and broke...

.1 Yes, all but one of the managed systems have been assigned an existing scan.

538.0 Most scan and event names and references to them are in upper case, we
also tend to use the DECwindows interface into the PCM editor which should mean
that case mismatches disappear, but it's a tip worth remembering.

Shortly before he left (about ten days ago) Steve applied MUP1, and we both
edited the profile. Shortly afterwards we got complaints about no messages in
multi-line windows, except for those of the Console Manager class.

The fragment of DAEMON flagged controller process log in .0 shows that the
assigned scan exists but has no events in it. This is true, it has other scans
in it. Other parts of these log files show that systems assigned the scan HSJ40
were not being reassigned the NULL scan, they had 5 events. The scan HSJ40 has 5
events in it, and also a couple of scans which should increase the number of
events associated with the system.

It would appear that the recursive nature of scans has disappeared. Is this
intentional? How do I get it back?

    No, the recursive nature of the scan profiles has not been removed.
    I don't know what changes were made to the MUP concerning this, but I 
    suspect that there were little it any.
    
    Perhaps the best option now is for you to 'export' your database and
    post it as a reply. We can then load your profile and see if we cn come
    up with anything. 
    
    Just for kicks, you could create a new scan profile, called HSJ40_NEW
    and readd the events and then reassign it to one of your systems and 
    see what happens. (Don't forget to reconfigure after making changes).
    
    Also, is this happening for all systems or is it just for the HSJ40?
    
    Dave

    oh boy.... we have run into the same thing here at a customer site and
    we have reproduced it using their profile. It's not a matter of
    recursion but it appears that events in "nested" profiles will no
    longer trigger. You can create a profile and define some events that
    exist *only* in that profile; assign that profile to a node and you
    can now trigger those events at will. Once they are nested within a
    sub-profile they cease to function.
    
    Regards,
       Dan

  Hmm, there were NO specific changes in the MUP around events or nested
  scans, however, I guess one of the other changes we made broke something
  somewhere!

  We are on holiday in the UK until the 3rd, I will take a look at this as 
  soon as I get back in the office!!

  I guess in the meantime, the workaround is to not use nested scans.

Cheers,
Phil

  OK, OK I own up, I made a mistake!!! It would appear that when I added some 
  code to make sure that we didnt have duplicate recursive scans, I used a !=
  instead of an == in an if statement, so, instead if not getting duplicate
  recursive scans you get NO recursive scans at all!!!!! 

  As soon as I have tested the fix I will make a new image available next 
  week when I return to work from my Christmas break.

Cheers,
Phil

Well - I'm glad we have discovered the problem before too many customers
have applied the MUP.  My concern is that a number of our customers would
consider this issue as a show stopper since they rely so heavily on event
notification.  Has the MUP shipped, or can we catch it before it goes out
the door from SSB?  

Please advise.

Connie

Connie,

 The MUP has been withdrawn, the only customers who will have it are those 
 who have been given it by readers of this conference.

 We will be resubmitting the kit tomorrow with the necessary fix included.

Cheers,
Phil