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

Conference humane::scheduler

Title:SCHEDULER
Notice:Welcome to the Scheduler Conference on node HUMANEril
Moderator:RUMOR::FALEK
Created:Sat Mar 20 1993
Last Modified:Tue Jun 03 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1240
Total number of notes:5017

1239.0. "HOLDing - Custom Load Balance" by TAVEIS::LANDAU (DEC Israel IM&T) Mon Jun 02 1997 05:42

        DECscheduler question -
        Some jobs are going into HOLD state right after the machine
        startup.
        It's related to Load Balancing; the job owner gets the following
        mail message: HOLDing - Custom Load Balance.
        Why does it happen and how can I fix it?
    
        This is the log file:
        Nsched Version V2.1B-7  starting...
        Setting Debugging OFF
        Setting Job Max to  10
        Setting Logging to  5
        Setting Default Job Priority to  4
        Setting Load Balancing ON
        Setting Restart Params to CLEAR on job completition
        Setting Remote Jobs ENABLED
        Setting brkthru/notify wait to  300  seconds
        CPUtype= 19  CPU_count= 1  Total pages= 508560   Meg= 260   VUPS=
    
        Thanks in advance.
        Cross-posted in the POLYCENTER notesfile.
T.RTitleUserPersonal
Name		DateLines
1239.1infoRUMOR::FALEKex-TU58 KingMon Jun 02 1997 18:3818
    There is a logical NSCHED$CUSTOM_LOAD_BALANCE (or something similar),
    that can turn on the custom load balancing facility on/off. When set, the
    default scheduler spawns a command procedure that makes the decision of
    where a job should run, rather than picking the node with the highest
    service rating.  The spawned procedure is called nsched$:lbal.com or
    load_balance.com or something similar (see documentation). It receives
    information about the candidate job and at the end should
    define a logical name to tell the scheduler which node to pick.
    
    Check if this logical is set by doing $ show logi nsched$*
    The logical is probably being set in the scheduler's start up
    procedure (in sys$startup:).  If so, look at the file that is executed
    and see if that's what's holding the jobs/sending mail.   If you don't
    want custom load balancing you should shut it off by deassigning the
    logical.  Only the one on the "default" scheduler is looked at, but
    settings should be the same on each node in case the "default" role
    moves.