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

Conference csc32::consolemanager

Title:POLYCENTER Console Manager
Notice:Kits, Scans, Docs on CSC32:: as PCM$KITS:,PCM$DOCS:, PCM$SCANS:
Moderator:CSC32::BUTTERWORTH
Created:Thu Aug 06 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1541
Total number of notes:6564

550.0. "Controler ACCESS VIOLATE and Byte quota?" by ULYSSE::WEISSBECK () Thu Jan 12 1995 10:22

    What are the correct process quota to avoid the Controler process to
    crash:                                
    
    $ !
    $ ! Start a Child Controller process, number 1
    $ !
    $ CHILD :== $CONSOLE$IMAGE:CONSOLE$DAEMON.EXE
    $ CHILD
    POLYCENTER Console Manager
    Console Controller Daemon Version V1.5-006
    
    CMProcessLocalAccept: failed to create output mbx!!
    Failed to create Local socket CONSOLE_CTRL_GORBIO
    exceeded byte count quota    <----------------------------------- ????
    %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual
    address=0000009B, PC
    =00008DDC, PSL=03C00005
    
      Improperly handled condition, image exit forced.
    
            Signal arguments              Stack contents
     
            Number = 00000005                003D98A0
            Name   = 0000000C                003D9801
                     00000000                00332F90
                     0000009B                0034AEC0
                     00008DDC                00000000
                     03C00005                201C0000
                                             7FE9B500
                                             7FE9B4E0
                                             00019405
                                             00008B9C
    
            Register dump
            R0 = FFFFFFF8  R1 = 8684FA50  R2 = 00000000  R3 = 003D98A0
            R4 = 00000000  R5 = 0034AEC0  R6 = 00019C7C  R7 = 00000000
            R8 = 00000000  R9 = 00000000  R10= 00000000  R11= 00000000
             AP = 7FE9B45C  FP = 7FE9B41C  SP = 7FE9B498  PC = 00008DDC
            PSL= 03C00005
    
    $ EXIT $STATUS
       SYSTEM       job terminated at 12-JAN-1995 10:45:15.08
    
      Accounting information:
      Buffered I/O count:             580         Peak working set size:2153
      Direct I/O count:               278         Peak page file size:  10798
    
T.RTitleUserPersonal
Name
DateLines
550.1CSC32::BUTTERWORTHGun Control is a steady hand.Thu Jan 12 1995 21:5026
    There is no absolute answer for your question.
    What are the settings of SYSGEN params DEFMBXBUFQUO DEFMBMXMSG?
    Chances are the setting of DEFMBXBUFQUO is rather high and every
    temporary mailbox that isn't created with hardcoded values will use
    the sysgen params above.
    
    You can also tune the parameters for the child daemons with the
    following magic logicals. Note that the values listed are the defaults
    for each quota. Youwill not find these logicals defined by default:
    
    
          CONSOLE$DAEMON_ASTLM               1024
          CONSOLE$DAEMON_BIOLM               1024
          CONSOLE$DAEMON_BYTLM             327680
          CONSOLE$DAEMON_CPULM                  0
          CONSOLE$DAEMON_DIOLM               1024
          CONSOLE$DAEMON_FILLM               1024
          CONSOLE$DAEMON_PGFLQUOTA          30000
          CONSOLE$DAEMON_PRCLM                 64
          CONSOLE$DAEMON_TQELM               1024
          CONSOLE$DAEMON_WSQUOTA             2048
          CONSOLE$DAEMON_WSDEFAULT            512
          CONSOLE$DAEMON_ENQLM               1024
          CONSOLE$DAEMON_WSEXTENT            8192
          CONSOLE$DAEMON_JTQUOTA             1024
    
550.2DEFMBX... Quotas responseELANSE::PLENTFri Jan 20 1995 14:3211

	The two parameters are set with the following values:

	DEFMBXBUFQUO=2112
	DEFMBMXMSG=256

	Note that if I increase DEFMBXBUFQUO the problem is delayed but still 
	occurs !

	Thanks for your help
550.3CSC32::BUTTERWORTHGun Control is a steady hand.Tue Jan 24 1995 19:5410
If you increase the value of DEFMBXBUFQUO then your allocating even more
    bytlm per mailbox. Try this:
    
    DEFINE/TABLE=CONSOLE$LOGICAL_NAMES CONSOLE$DAEMON_BYTLM 80000    
    
    This will double the available bytlm to the child daemons. The value
    I posted in -2   of 327680 is wrongs. It's really 32768!
    
    Regs,
      Dan
550.4CONSOLE$DAEMON_BYTLM: does it have any effect?VELI::KORKKOVeli K�rkk�, FNO/OMS Delivery, DTN 879-5512Sun Feb 12 1995 17:319
        Re .-1
        
        Running PCM V1.5 MUP A. To me it appears that the Daemons have
        BYTLM quota around 600 000 even without any logical names. The
        values appears to be as if it was inherited from process
        "Console Daemon". I hope those Ctrl processes are not "pooling"
        the BYTLM with each other and Daemon?
        
        veli
550.5OPG::PHILIPAnd through the square window...Mon Feb 13 1995 10:5610
Veli,

  The child daemons are created as detached processes, they do not inherit
  any values from thier parent, the creprc call explicitly sets the quota
  values, when you define these logical names, it is the creprc values that
  you are changing. Now, having said that, we dont allow you to lower the
  values with the logical names, only increase them.

Cheers,
Phil
550.6BUFIO byte count/limit 571894/573942 for Console Ctrl 01VELI::KORKKOVeli K�rkk�, FNO/OMS Delivery, DTN 879-5512Mon Feb 13 1995 17:3028
So, reading all replies in this thread, do you expect following:

Process index: 0021   Name: Console Ctrl 01   Extended PID: 00014321
--------------------------------------------------------------------
Status : 00140001 res,phdres,login
Status2: 00000001 quantum_resched
PCB address              8425BC00    JIB address              84182180
PHD address              89329800    Swapfile disk address    00000000
Master internal PID      01430021    Subprocess count                0
Internal PID             01430021    Creator internal PID     00000000
Extended PID             00014321    Creator extended PID     00000000
State                       HIB      Termination mailbox          0000
Current priority                9    AST's enabled                KESU
Base priority                   4    AST's active                 NONE
UIC                [00001,000004]    AST's remaining               971
Mutex count                     0    Buffered I/O count/limit     1006/1024
Waiting EF cluster              0    Direct I/O count/limit       1008/1024
Starting wait time       1B001B1B    BUFIO byte count/limit     571894/573942
Event flag wait mask     DFFFFFFF    # open files allowed left     974
Local EF cluster 0       60400000    Timer entries allowed left   1005
Local EF cluster 1       C0000000    Active page table count         0
Global cluster 2 pointer 00000000    Process WS page count        3164
Global cluster 3 pointer 00000000    Global WS page count          204

I do not have any CONSOLE$DAEMON_xxxx logicals defined? According to earlier
replies, I would not have expected to see so high BYTLM there?

Veli
550.7CSC32::BUTTERWORTHGun Control is a steady hand.Tue Feb 14 1995 01:495
    No we don't. The two possibilities are that the logicals are defined or
    the sysgen param PQL_MBYTLM has been increased to 600000.
    
    Regs,
      Dan
550.8OPG::PHILIPAnd through the square window...Tue Feb 14 1995 09:306
Veli,

  Is this causing you a problem? If not, why are you so concerned?

Cheers,
Phil
550.9no problems, my PCM just does play according the rules ....VELI::KORKKOVeli K�rkk�, FNO/OMS Delivery, DTN 879-5512Tue Feb 14 1995 22:4518
>  Is this causing you a problem? If not, why are you so concerned?
        
        I have had the same RWMBX problems and our external/internal  business
        relies quite heavily now on PCM. I've been investigating these
        RWMBX problems, read these notes very carefully, tried to
        implement the ideas if they seem reasonable. As part of
        "doublechecking" things just happened to note that Console Ctrl
        0x processes had already QUITE HIGH Bytlm quotas.
        
        We do not have ANY logicals of type CONSOLE$DAEMON... and
        neither can this behaviour (large BYTLMs ) be because of PQL
        params. PQM_xBYTLM are considerably lowe than 570 000.
        
        So, on my system PCM does not behave the way it should according
        to this string of replies. Hopefully this "bug" will not fixed
        on possible future MUP kit :-)
        
        veli
550.10OPG::PHILIPAnd through the square window...Wed Feb 15 1995 10:428
Veli,

  The RWMBX problem has been IPMT'd and I am currently spending my
  time fixing this problem, you should see something you can test
  in a couple of weeks hopefully.

Cheers,
Phil