Conference pamsrc::decmessageq

    
    The service dependency only guarantees that one service
    starts before another.  The MessageQ service  needs  to
    start up the groups and that can take time  (especially
    when there are other things coming up at boot time).
    
    _sjz.

    
    
    Hmmmm... Do you mean that
    
    1) Service Control just orders the startup list by the dependancies??
    It does not REALLY wait until a service has reported a state of
    STARTED as opposed to STARTING
    (which I take to mean: Still initializing stuff)
    before kicking off other services with dependancies?
    
    				OR
    
    2) Does DMQ not wait till it has started it's groups to report a STARTED
    state to SCM???
    
    I had to fire off threads just to talk to SCM while I verified that I
    could access and initialize my  DMQ and WinSock environment before
    I reported STARTED to SCM.  If the answer is #1, THAT SUCKS.
    Why bother to define the intermediate SCM states!
    
    If the answer is #2) Well....um... Why?
    And more importantly, whats the best way to handle it? Loop and wait?
    
    
    Thanks,
    Garrison

    
    Hi,
    
    Any idea which scenario (from .-1) explains the observed behaviour,
    and the best way to fix or work around it.
    
    If I just keep trying to attach, will it just work at some point, 
    or is my session screwed once it fails??
    
    My customer is waiting for an explanation of why this is occuring,
    and how we will fix it.
    
    
    thanks again,
    Garrison

    Well, so far, not so good...
    
    I attempted to wait and then retry my pams_attach_q call in a loop
    while I got PAMS__RESRCFAIL as a return code.  After about 3 seconds
    (~90 retrys) I got:
    
    	PAMS__PAMSDOWN, DECmessageQ group is down - COM Server has exited.
    
    Any feed back for a work around would be GREATLY
    appreciated.  Im shooting blind here, I don't know if this retry method
    is doomed or if I just need to tweak my wait time between attempts or
    ?....
    Thanks in advance for any suggestions!
    Garrison

re: .4

>    I attempted to wait and then retry my pams_attach_q call in a loop
>    while I got PAMS__RESRCFAIL as a return code.  After about 3 seconds
>    (~90 retrys) I got:
    
>    	PAMS__PAMSDOWN, DECmessageQ group is down - COM Server has exited.

	  From this it looks like the group went down.  Please take a look
	and see if it went down and whether the reason was logged.



	Marty

    Here are the logs. CcsCronus is our service. I changed the wait before
    retry on the pams_attach_q to 1 second (from 100ms).
    
    1997-05-28 09:04:18  CcsCronus starting on BarNoneNT
    1997-05-28 09:04:19  pams_attach_q Failed with PAMS__RESRCFAIL.
    1997-05-28 09:04:20  pams_attach_q Failed with PAMS__RESRCFAIL.
    1997-05-28 09:04:21  pams_attach_q Failed with PAMS__RESRCFAIL.
    1997-05-28 09:04:22  pams_attach_q Failed with PAMS__RESRCFAIL.
    1997-05-28 09:04:23  DMQ Error - pams_attach_q
    1997-05-28 09:04:23  PAMS__PAMSDOWN, DECmessageQ group is down - COM
    Server has exited
    
    
    dmqgcp.exe (143) 28-MAY-1997 09:04:22 **********
    gcp, group control process for group 9 is running
    
    dmqqe.exe (146) 28-MAY-1997 09:04:23 **********
    qe, queuing engine is running
    
    dmqld.exe (154.155) 28-MAY-1997 09:04:25 **********
    ld, link listener for group 9 is running
    
    dmqld.exe (157.150) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 1 is running
    
    dmqld.exe (111.107) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 2 is running
    
    dmqld.exe (111.107) 28-MAY-1997 09:04:25 **********
    ld, operation failed to complete
    
    dmqld.exe (111.107) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 2 is exiting
    
    dmqld.exe (111.159) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 3 is running
       
    dmqld.exe (149.160) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 4 is running
       
    dmqld.exe (149.160) 28-MAY-1997 09:04:25 **********
    ld, operation failed to complete
    
    dmqld.exe (149.160) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 4 is exiting
    
    dmqld.exe (110.159) 28-MAY-1997 09:04:25 **********
    ld, operation failed to complete
    
    dmqld.exe (110.159) 28-MAY-1997 09:04:25 **********
    ld, link sender for group 9 to group 3 is exiting
      
    
    Let me know if there is any other logfile or trace data that would
    help.
    
    Thanks!
    Garrison
       

    
    try doing a pams_exit() between each failed attach until
    your initial successful attach.
    
    _sjz.