Conference iosg::all-in-1_v30

    I have found other %SYSTEM-F-IVLOCKID, error messages such as :
    
     7-JUN-1993 09:18:59.24  Server: CNB06V::"73="  Error:
    %SYSTEM-F-IVLOCKID, invalid lock id  Message: CsiCacheBlockAstService;
    Error releasing VMS lock
    
     7-JUN-1993 09:34:45.00  Server: CNB06V::"73="  Error:
    %OAFC-W-BUFFOVFL,
    Client buffer not big enough for requested information  Message:
    CsiEndFunction; Error putting EOC on out stream
    
    Quite a few of these around.
    
    Sunil

    Hi Sunil,
    
    The "insufficient privilege or file protection violation" error
    suggests that the DOCDB file does not have system access on it. This
    would mean that ALL-IN-1 and the FCS would not be able to open it,
    because they do not raise BYPASS privilege. 
    
    Cheers
    
    Stuart

    Hi Stuart me old china,              
    
    Re: .2
    
    >-< No system access? >-
    
    That's what makes it one of those hair pulling problems.  I had checked
    this and system has RWED protection set and so do all the datafiles,
    that's why I am stuck.  When you do a read from the Manage Drawer menu
    of the drawer it shows it as not a shared drawer, if you try and edit
    drawer access it is shared !!!!
    
    Regards,
    
    Sunil
    
    PS - for all of you out there "Me old china" means mate.  China -> plate
    -> mate rhyming slang from London :-).

    Hi Sunil,
    
    The drawer shared flag is determined by whether or not there is an ACL
    on the drawer's ACCESS.DAT file. So do a $dir/sec of it and see whether
    there is one.
    
    Are logical names getting in the way? Does the drawer directory name
    contain a logical name. If so see if there are multiple definitions of
    the logical.
    
    Cheers
    
    Stuart

    Hi Stuart and Paul,                                 
    
    Re: .4
    
    >The drawer shared flag is determined by whether or not there is an
    >ACL on the drawer's ACCESS.DAT file. So do a $dir/sec of it and see
    >whether there is one.
    
    That was the first thing I had checked when the call was opened and the
    ACL's were present.  Also the logical names was checked, well we told
    the customer while they were still in the testing phase of ALL-IN-1
    months ago and they don't use them.  I should have written all of this
    down sorry.
    
    Re: .5
    
    >Have you checked that the drawer directory(s) have  S:RWE access?
     
    Yep it's set to S:rwed.
    
    >Maybe you shopuld look over the PARTITION record for the drawer and see
    >exactly what it says.
     
    I had done this and all appeared to be fine.
    
    Last night I had told the customer to shutdown the server and restart. 
    The customer did this and the drawer is now a shared drawer, they
    didn't change anything.  The logfile contained a number of errors but
    most of them were the  "Message: CHANNELCNT low - flushing drawer from
    cache" messages.  Followed by the "Message: Drawer flushed from cache for 
    IOCHANNELCNT", which is quite normal from what has been said/discussed
    in the past in here.  This message was written to the logfile when we
    were having problems with the drawer.
    
    Later on in the evening we got the following error messages:
    
    12-AUG-1993 22:51:36.90  Server: CNB08V::"73="  Error:
    %MCC-E-EXISTENCE_ERROR, object does not exist 
    
    This has not caused any problems, I would like to know what it means.
    
    My plan of action I have put forward to the customer is to carry out
    further tuning exercise for the FCS.  The customer has been quite good
    and monitors the system and has said that they try to maintain an
    optimum environment as per the manual.
    
    The shuting down the server to solve the shared drawer problem seems a
    bit to harsh a measure to solve the problem.  From what Bob has said in
    the past it would appear that this should never be required.  Have any
    of you got comments on this ?  If I can get more information please let
    me know, what is required and I'll see what I can do.
    
    Regards,
    
    Sunil
    
    PS - Paul I hear that you are going to work for the company that I was
    working for in England.  I wish you all the best thanks for all your
    help.

    Hi Sunil,
    
    I suspect that the ACL handling in FCS is a bit suspect. It's one area
    which I paid some attention to recently, so in the event of their being
    a future patch release of FCS, you might find some improvement.
    
    Patience, my friend, is all I can suggest!
    
    The MCC error simply states that an invalid handle was passed to an MCC
    routine to operate on. Could be a synchronisation problems or memory
    corruption, or, or, or ...?
    
    Like they say - don't worry about what you can't control.
    
    Paul.

    Hi Trev,
    
    >provide a better more lasting fix/workaround?            
    
    My customer is using multiple servers and has not complained since.  So
    far there are a couple of large sites that had FCS problems making use
    of multiple servers currently and they appear to be quite happy.
    
    Multiple servers has been discussed in topic 3137, this maybe of some
    help to you.
    
    Regards,
    
    Sunil