Conference azur::mcc

6308.0. "Insufficient event pool memory revisited" by BUDDIE::KENWORTHY () Mon May 22 1995 22:31

    I have a customer that is having a problem similar to the one discussed
    in note 5878.
    
    This customer is running the following:
    	DECmcc 1.4
    	Ultrix 4.3
    
    They are using the ip_poller to poll approx. 300 nodes every 90 seconds
    (performance is awful)
    
    They are getting the following failure when running a "do" command from
    the command line.
    
    Failure in Kernel event reciept. Thread terminating:
    %MCC-E-INSEVTPOLMEM, Insufficient event pool memory to post event.
    
    I suggested to the customer that they increase the variables that are
    mentioned in note 5878, ie MCC_EVENT_POOL_SIZE 2097152 and
    MCC_EXENT_EDQ_SIZE_LIMIT 2500. The customer put the variables into the
    mcc_login.csh file.
    
    When the customer ran MCC the following error occurred:
    %MCC-E-EVT_POOL_SIZESKEW Requested event pool size incompatable with
    existing pool. They then get an IOT trap and MCC dies.
    
    What are we doing wrong? Is there anything else we need to do or are
    there other dependencies that need to be checked?
    
    Thanks in advance.
    regards
    jimk

Jim,

The SIZESKEW error arises when a process tries to create the event pool with one
size, while there are still processes using it with the old size. 

If you change the size of the event manager, this change must appply to all MCC
processes and all MCC processes must be stopped. As you are running DECmcc on
Ultrix, don't forget that management modules stay around as separate processes
even if the PMs have been stopped. These processes should be stopped also. 

That should clear the SIZESKEW problem.

Best regards,

Marc. 

    The customer is NOT running DECmcc version 1.4. They are running DECmcc
    version 1.3
    
    regards
    jimk
    

Marc's reply was valid for both versions.

Benoit