Conference csc32::consolemanager

Jim,

  I dont want to sound unhelpful, but, if you really need an urgent
  fix for this problem then you will have to escalate it in the IPMT
  system, we dont have the time that is required to fix this at the
  moment, however, an IPMT will force us to make time (it will also
  highlight to our management that this "design deficiency" is quite
  important, and requires time to work on it).

  We as you say are having difficulties reproducing this problem,
  however, we have seen enough systems where this has occured to
  understand the problem and we have formulated a strategy to fix it,
  its just a matter of scheduling it in with our other work. Now, as
  I have said, if you IPMT it, then it will move up our work queue
  to the top (or extremely close to it).

  This goes for anyone else who is experiencing this problem, IPMT it
  and we will work on it, until that is done unfortunately we cant do
  anything.

  This all boils down to the fact that this NOTES file IS NOT an official
  problem reporting mechanism.

Cheers,
Phil

    I had these problems and I resolved them.
    
    PCM is used to manage 12 nodes VMS but I need to open today 24 windows
    with the action console$eventlist.
    
    I had always processes in RWMBX or RWAST, ACCVIO ...
    The problems has been resolved by:
    
    	- defmbxbufquo	: 1054*4
    	- defmbxmxmsg	: 256*4
    
    	- bytlm		: 800000	! Proc console$startup modified
    	- diolm		: 256		! Proc console$startup modified
    
    If it can help somebody.

    Re. -1.
    
    Your input is appreciated but it's still possible to cause an
    "RWMBX hang". Your tuning makes it much less likely and is quite
    appropriate for a lot of environments. The problem is a design flaw - 
    I've spent many hours in the code looking for a simple fix and I can
    tell you that  it doesn't exist - and it will have to be corrected by
    some rather major code changes. The code that needs changing is very
    basic to the product and it will require some exhaustive testing
    to make sure we don't break anything else.
    
    Regards,
       Dan

		Re. -2.


	In CONSOLE$STARTUP.COM You specify to modify CONSOLE$STARTUP.COM
	the BYTLM and DIOLM value. With the CONSOLEMUP015 version however
	i cannot find these parameters in the procedure.

	Could it be You mean to add/modify /IO_BUFFERED and /BUFFER_LIMIT
	to the RUN command instead of using the PQL/other values?


	Thanks on forehand,
	Erik.

    > Could it be You mean to add/modify /IO_BUFFERED and /BUFFER_LIMIT
    
    /BUFFER_LIMIT is indeed equivalent to BYTLM in the PQL's and UAF.
    and /IO_DIRECT is equivalent to DIOLM.
    
    /IO_DIRECT is not specified in the command file anywhere so we are
    picking up the PQL sysgen value here. Note also that the buffer limit
    value is passed as an argument to the CONSOLE$IMAGE_STARTUP DCL
    subroutine in the CONSOLE$STARTUP procedure. The MUP specifies
    1024000 for the ENS daemon and 655530 for the parent controller daemon.
    Note that the values that you see in CONSOLE$STARTUP *do not* effect
    the child controller daemons. The quotas for these processes are
    either defined by a logical ( described in another recent note) or
    a default value contained within the image.
    
    Regs,
      Dan
    
    
    
    

    Have you got an idea when this "design flaw" will be fixed.
    
    We have scheduled a major upgrade weekend in a couple of weeks, where
    these RWMBX problems are absolutely not acceptable.
    
    Is this issue "high" on the priority list?
    
    Marc

Marc,

>>    Have you got an idea when this "design flaw" will be fixed.
  
 When we get the time.
  
>>    We have scheduled a major upgrade weekend in a couple of weeks, where
>>    these RWMBX problems are absolutely not acceptable.

 Yes, we realize these problems are not acceptable and they will be fixed.
    
>>    Is this issue "high" on the priority list?

 Yup, it number 1 but, as I said its just finding the time to schedule it in.
    
Cheers,
Phil

    Hi,
    
    Is there any news regarding the RWMBX-poblems? We will be using
    PCM extensively soon during the scheduled upgrade weekend.
    
    Could you give us an idea when this issue will be resolved via
    patch/MUP?                                    
    
    Thanks
    
    Marc
    

Marc,

  Nobody has IPMT'd the problem yet, until it is officially elevated, we cant 
  work on it until we have fixed all those officially elevated problems.

  So, dont hold your breath.

Cheers,
Phil

    Ok,
    
    I will stick with the rules. I have logged a new call at our CS support
    center, to IPMT the issue officially.
    
    Marc

    Just to revisit this issue....
    
    I am now working with a large customer who is nearly ready to go live
    with PCM and who has experienced this problem regularly during testing.
    
    Since it has been about a month since the IPMT was issued, is there any
    additional information on when a fix might be available?
    
    Thanks in advance,
    
    Larry

Larry,

  I am busy trying to get the software working correctly again after 
  putting the fix in (we have a few minor teething problems), we will be
  issuing IFT test kits in about two weeks time if we can get all the present
  problems fixed.

  We will not be releasing a MUP or ECO for this fix, it is too large a 
  change in the code for that, so, we will be doing a small field test with 
  those people who have experienced this problem, then we will be releasing 
  V1.6 of PCM.

  A warning to everybody who is reading this...

  The underlying Interprocess Communication mechanism has completely changed, 
  any user written action routines using the API MUST be recompiled and link in 
  order for them to work with this new version. This applies ONLY to OpenVMS 
  (VAX amd Alpha).

  Finally, we will be taking this opportunity (V1.6) to release a brand new
  C3, this new C3 is hierarchical in nature. Also, we WILL NOT be releasing
  V1.6 on ULTRIX, any customers wishing to get the new C3 MUST upgrade to 
  either Digital UNIX or OpenVMS.

Cheers,
Phil

        We would be very interested to FT your coming V1.6 kit. 
        
        veli

    
    
    We are interested to participate in the FT.
    
    /Staffan

    
    We are also interested!
    
    
    /Gianmario
    /Paolo

     If this is going into EFT, then my customer would like to Participate
    
    				Steve C

Do you have any idea when the EFT version will be available?

Johan

Johan,

  When it works!!! Hopefully next week sometime, but no promises.

Cheers,
Phil

    
    ...so you'll let us know then...
    
:-)
    
Nige

  Of course.

Cheers,
Phil

The computer center operations manager at CERN has requested to field test
the next version.

How can I apply for a formal or informal FT?

-cw

cw,

  We are only going to be doing this informally, so just keep an eye open
  here and grab the kit when it becomes available.

Cheers,
Phil

    Any progress on the 1.6 EFT or SSB kit ?????
    
    
    		Steve C

Steve,

  Any day now, we are ironing out the final few obvious problems.

Cheers,
Phil

    
    Well we hung it out for a while but our use probably doesn't
    need the might of a console manager so I'll take my leave now as
    reverse LAT's probably all we really need. All the best for the new
    version, I may come back to you if we can't do all we want.
    
Nige