[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference noted::sns

Title:POLYCENTER System Watchdog for VMS OSF/1 ULTRIX HP-UX AIX SunOS
Notice:Wishes:406,FAQ:845,Kits-VMS:1000,UNIX:694 VMS ECO01 FT kit: 521
Moderator:AZUR::HUREZZ
Created:Fri May 15 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1033
Total number of notes:4584

992.0. "Process Hang Detection?" by HGOVC::PATRICKNG () Wed Feb 05 1997 04:13

    Hi all,
    
     I would like to know whether Polycenter Watchdog can detect the
     following scenairo :
        
     A program send and receive to and from the mailbox. This program
     will hang if the mailbox is full or other reasons. System people do not
     know the problem until end-users compliant the application hangs.
        
     Is Watchdog able to detect the process hang? How to achieve this
     goal.
    
    Thanks,
    Patrick
T.RTitleUserPersonal
Name		DateLines
992.1Not (yet) built in...AZUR::HUREZConnectivity & Computing Services @VBE. DTN 828-5159Wed Feb 05 1997 08:2223
    The watchdog isn't able to detect such hanging upon I/O situations,
    unless the considered process performs busy waiting and therefore
    eats CPU, which would be detected as a process looping event.  Such
    hanging situations as you describe may be normal (process is waiting
    for another one to read the mailbox and free space for the new message
    into it), or the programmer would have added the IO$M_NORSWAIT flag
    to the $QIO Write operation, in order to avoid the hang.  However,
    it could be interesting indeed to add an event check for this into the
    Agent, for given processes that could have been badly programmed.
    
    In waiting, one could use the
    	SNS$EXAMPLES:SNS$CHECK_PROCESSES_STATES.COM
    sample command file that will detect process stuck in odd states
    (essentially MWAIT state, but possibly one of the following:
    MWAIT/CEF/COM/COMO/CUR/COLPG/FPG/HIB/HIBO/LEF/LEFO/PFW/SUSP/SUSP0)
    
    You may get some help about how to run it, using:
    	@SNS$EXAMPLES:SNS$CHECK_PROCESSES_STATES.COM HELP
    
    Regards,
    
    	-- Olivier.
992.2VELI::KORKKOVeli K�rkk� @FNO, 879-5512Sun Feb 16 1997 09:3911
        SNS$CHECK_PROCESSES_STATES.COM seems to have a small problem. At
        least on OpenVMS V7.1, f$pid(0) is not allowed. If I substitute
        
        	$ ctx=""
        	... f$pid(ctx)...
        
        instead, it seems to work as intended.
        
        _veli
        
        P.S. Yeah, V7.1 is not supported SNS-wise.
992.3Fixed.AZUR::HUREZConnectivity & Computing Services @VBE. DTN 828-5159Mon Feb 17 1997 08:534
    Thanks, Veli;  This is now corrected in the CMS,
    ready for next ECO/release...
    
    	-- Olivier.