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

Conference iosg::all-in-1_v30

Title:*OLD* ALL-IN-1 (tm) Support Conference
Notice:Closed - See Note 4331.l to move to IOSG::ALL-IN-1
Moderator:IOSG::PYE
Created:Thu Jan 30 1992
Last Modified:Tue Jan 23 1996
Last Successful Update:Fri Jun 06 1997
Number of topics:4343
Total number of notes:18308

3735.0. "Filecabinet server and SYSUAF.DAT" by ANGLIN::HARRISA (lost in the jungle of doubt) Wed Jan 05 1994 17:40

    ALL-IN-1 3.0-1, VMS 5.5-2
    
    the filecabinet server on a cluster is accessing the SYSUAF.DAT file.
    i beleive this is normal. when thecustomer uses a home grown procedure,
    they're having trouble access the SYSUAF at times. the VMS manager
    thinks thr problem is probably in their procedure.
    
    i just want to make sure that the filecabinet server is doint the right
    thing here.
    
    	ann
    
T.RTitleUserPersonal
Name
DateLines
3735.1FCS should be accessing itCHRLIE::HUSTONWed Jan 05 1994 19:557
    the FCS opens sysuaf at startup time and reads/writes to it during
    connection to the FCS, both from TL and from IOS.
    
    What problems are they having accessing it?
    
    --Bob
    
3735.2server errorANGLIN::HARRISAlost in the jungle of doubtFri Jan 14 1994 15:3814
     re.1, the problems are with a home grown procedure. the customer has a
    workaround for their procedure. 
    
    same system, they had to stop the FCS processes on 2 nodes in the
    cluster and restart the servers. the error message in the server log
    is:
    
    ERROR:MCC-E-IN_USE_ERROR, in use error 
    Message:CsiCacheFlushDrawerAccess; Error from mcc_mutex_try_lock
    
    what does this mean?
    
    	thansk - ann
    
3735.3is this possible?ANGLIN::HARRISAlost in the jungle of doubtThu Feb 10 1994 23:0830
    ALl-IN-1 3.0-1 VMS 5.5-2
    
    ok - now the customer wants a solution.
    
    they have a program called "rights". rights is used to grant/remove
    identifiers on users for accessing various business applicatins.
    
    my customer is able to use his rights program to grant or remove
    identifiers, but they are not able to just LOOK at who has what
    identifiers.
    
    they can stop the FCS, but have to call me (ALL-IN-1 manager) to
    restart it. the concerns are: 
    
    	1. what harm will stopping and starting the FCD several times
    during the day cause?
    
    	2. is it possible to run the FSC startup procedure outside of the
    ALL-IN-1 menus? i was thinking of a .COM file that went into ALL-IN-1
    manager account with/NOINIT and ran the FCS startup.  but i'm having
    trouble setting the FCS symbols. when i run the
    OAFC$SERVER_STARTUP.SCP;1 from /NOINIT, it keeps saying that the server
    is remote.  my test system is a 2-node cluster. the problem site is a
    3-node cluster.
    
    thanks for any thoughts/advice.
    
    	ann
    
    
3735.4� an answerIOSG::STANDAGEFri Feb 11 1994 07:5322
    
    Ann,
    
    The impact of stopping the server very much depends on the nature of
    the ALL-IN-1 environment. If the site uses TeamLinks at all then the
    impact will be immense, such users will obviously not be able to
    operate within their ALL-IN-1 filecabinet.
    
    Other operations between drawers will not work either e.g. copying a
    document between drawers, etc. Possible future versions of ALL-IN-1
    will interact more and more with the server, so your situation is
    likely to get worse ! 
    
    If you have the DSO option, and there are remote systems connecting to
    your drawers then you should consider this also. While your server is
    down, no remote sites can connect to your drawers.
    
    
    Kevin.
    
    
    
3735.5No theoretical effect, just unusableCHRLIE::HUSTONFri Feb 11 1994 18:2123
    
    re .3
    
    ann,
    
    Kevin gives a user point of view answer to the effect of stopping/
    starting the FCS. The only effect is that it will be unavailable
    for a period of time. Startup of the FCS is not a fast function.
    
    In reality, the effect will be nothing more than users cannot access
    anything that the FCS is required for: TL users, remote drawers, non-
    MAIN drawers, DSO options. When the FCS is shutdown (assuming it is
    done nicely, not with stop/proc) it allows all curently executing
    tasks to complete, so nothing will be lost.
    
    As for starting the FCS from outside IOS. Yes, possible, supported???
    I have done it on our cluster by feeding the correct parameters
    to the startup command procedure, note that this bypasses the 
    server data base that is within IOS, so I doubt it is supported
    (I had no choice, no system managers around at 11:00PM :-) )
    
    --Bob
    
3735.6GIDDAY::JOYCEBurn me kangaroo down sportMon Feb 14 1994 02:3118
    
    Re. -1
    
  >  Kevin gives a user point of view answer to the effect of stopping/
  >  starting the FCS. The only effect is that it will be unavailable
  >  for a period of time. Startup of the FCS is not a fast function.
  >  
  >  In reality, the effect will be nothing more than users cannot access
  >  anything that the FCS is required for: TL users, remote drawers, non-
  >  MAIN drawers, DSO options. When the FCS is shutdown (assuming it is
  >  done nicely, not with stop/proc) it allows all curently executing
  >  tasks to complete, so nothing will be lost.
   
    I have to disagree.  If a user is editing a document in another drawer
    in IOS and the FCS is not availble when they exit - then the document
    does not get unreserved.  Not a big deal I agree but it does cause
    concern to the users (bless 'em). 
    
3735.7IOSG::STANDAGEMon Feb 14 1994 09:3919
    
    >>I have to disagree.  If a user is editing a document in another drawer
    >>in IOS and the FCS is not availble when they exit - then the document
    >>does not get unreserved.  Not a big deal I agree but it does cause
    >>concern to the users (bless 'em). 
    
    Bob was refering to the server being stopped while it is actually being
    used. In your scenario the server is only being used
    during the initial create stage, and then the replace stage once the
    editing has been completed. In the middle of this there is no server
    operation. Even in this situation, if the user re-edits the document
    once the server comes back up, they regain the edited version, and can
    then successfully replace the document - so again, nothing has been
    'lost'.
    
    Kevin.
    
    
          
3735.8so close...ANGLIN::HARRISAsweet insanityThu Feb 17 1994 22:4620
    Thanks for the possible scenarios guys.  at first the customer told me
    that they were just going to stop the process at VMS level, i said NO
    NO !  call me at home and i'll stop it nicely.
    
    i am able to stop and start the server successfully from going in
    
    ALLIN1/NOINIT/USER=MANAGER
    OA$INI_INIT
     GET #OAFC_SERVER_NAME = $OAFC_SERVER_NAME
     DO OAFC$SERVER_STOP
     EXIT
    
    if i set the above up as a .COM file it doesn't work. if i type each
    line interactivly, i get the question " Are you sure you want to stop
    the server?" I answer Y and it proceeds.
    
    how can i preset the answer to Y ?
    
    	thanks - ann
    
3735.9UAF> show/rig /user=* works for me ?UTRTSC::SCHOLLAERTHolland goes USAFri Feb 18 1994 08:2612
    >my customer is able to use his rights program to grant or remove
    >identifiers, but they are not able to just LOOK at who has what
    >identifiers.
    
    Sorry for a late reply, but the FCS does not prevent 
    any program from reading SYSUAF... Try:
    
    UAF> show/rig /user=*
    
    Hope this helps,
    
    Jan
3735.10ANGLIN::HARRISAsweet insanityFri Feb 18 1994 15:4710
    Jan -
    
    .9 works fine, but i think the customers code might be a bit more
    involved. they told me they have about 40 hours of rewrites for it to
    correct this problem.
    
    the stopping/starting FCS .COM file is just a quick intermediate
    solution.
    
    	ann
3735.11ANGLIN::HARRISAsweet insanityFri Feb 18 1994 20:388
    ah duh... i modified the oaOAFC$SERVER_STOP.scp to bypass the "R U sure
    ?". I can now stop and start the server successfully outside of
    ALL-IN-1.
    
    if i think, it will happen!
    
    	ann