Conference iosg::all-in-1_v30

    
    What distribution level is the server running at (or trying to run
    at I should say)?
    
    What are you doing that is causing this. The server is already started, 
    then what do you do?
    
    What is happening is that you are doing something that is causing the
    server to try and add a partition to its cache (this means you 
    attempted to broker). The partition however is already in the
    cache.
    
    What version of the ALL-IN-1 and the FCS are you running. The versino
    of the FCS can be found in the startup message or anal/image on 
    sys$system:oafc$server.exe.
    
    --Bob
    

Bob,

>    What distribution level is the server running at (or trying to run
>    at I should say)?

To find this out, the server must be running.

>    What are you doing that is causing this. The server is already started,
>    then what do you do?

I'm trying to start the server, and it starts then immediately dies out.

At the moment any operation such as modifying the server's attributes, enable
server tracing, ect... requires that the server is running. So you can see, it's 
a vicious circle.

ALL-IN-1: ALL-IN-1 IOS Server for VMS V3.0 PBL123A  (US) ENGLISH 21-MAR-1992
FCS: "OAFC V1.0"

Is it possible to "reinitialize" the whole thing and start all over ?

Thanks for your time.

Lee.

    To re-initialize the server, delete the server parameters file
    (OA$DATA_SHARE:node$SERVERdecnet_object.DAT) and run the script
    OA$LIB:OAFC$SERVER_CREATE_73.SCP, modifying it if you are going to
    create anything other than a server for DECnet object 73 (or use
    OA$LIB:OAFC$SERVER_CREATE.SCP).
    
    Then try again...
    
    Tony

    
    re .all
    
    If you take the route Tony suggests (it will do what he says) please
    save the config file and send it to me, I would be curious about
    what is causing this.
    
    As for needing the server running, you are right, from the log message
    you included, it looked like the server was started. Sorry.
    
    Since you are trying to start the server, can you tell me the
    following:
    
    1) Are you on a cluster or standalone?
    2) If cluster, do you have a cluster alias
    3) If standalone, are SCSNODE and SCSSYSTEMID defined? If not what is 
       the value of OA$SCSSYSTEMID?
    
    During startup the server uses sys$getsyi to get the node ids of all
    the nodes in the cluster along with their names, these are then added
    to the server cache.  For the life of me I can't figure out how this
    would screw up and try to add the same name twice.
    
    I know you can't tell what distribution level the server is at since
    you can't start it up. They all come, by default, at level 0. Do you
    recall anyone setting it to 1? Could you send me your config file and
    i will see if my server can start it.
    
    --Bob
    

    Bob,
    
    1) We are on a standalone machine.
    2) VAXCLUSTER is set to 0
    3) SCSNODE and SCSSYSTEMID are defined.
    
    There are people who are trying to get Teamlinks to work on our system,
    so I can't assure you that the distribution level set to 0. 
    
    The next question is how do I reset it to 0 without having the server
    running ?
    
    Thanks,
    
    Lee.

    
    Lee,
    
    There is no way for TeamLinks to set the distribution level. It is done
    via the FCS system management interface and that is not available to
    non-VMS systems.
    
    Can you send me two things (via mail would be good):
    
    1) THe output of a $SHOW CLUSTER
    2) Your configuration file.
    
    As for resetting the distribution level, the best way is to wipe out
    the server record and re-create the server as was mentioned in an
    earlier note.
    
    If you give me your config file, I can try and start up my server with
    that. If that does not work, I can put a debug server on your node and
    go look at what is happening.
    
    --Bob
    

Bob,

The script OA$LIB:OA$PART_SEED.SCP was executed; I'm wondering if this has 
anything to do with the problem.

Files are sent.

Thanks,

Lee.

    
    Lee,
    
    Nope, the OA$PART_SEED.SCP would have no effect on this. That simply
    takes each user from the profile and adds their default drawer
    to the partition definition file (partition.dat).
    
    I used your config file to start my server and everything worked fine.
    You are at distribution level 0 and obj = 73 so it is neither of 
    those things.
    
    One thing occured to me, and I admit this is a shot in the dark. 
    You say you are standalone. Could is be possible that you have 
    the logical sys$cluster_node defined? We translate this at startup
    to get the cluster alias if present. I tested this on my system
    and if this is defined on a standalone system to be the node name
    then you would see the problem you're having.
    
    --Bob
    

Bob,

I don't believe this; just too vicious. This logical was defined for testing an 
application, and was only going to be around temporarily. 

I removed it and every thing is back to normal.

Thanks, I really appreciate your help.

Lee.

    
    Lee,
    
    Glad to hear it fixed it, I agree it is vicious and the server should
    be able to handle it. At that point we know you are not a cluster so
    we should be smart enough not to "second guess" ourselves, it will be
    fixed in one of the FCS baselevels going into EMERALD.
    
    --Bob
    

    Re .-1
    
    Bob really meant to say:
    
    	"This fix will be considered for inclusion in a possible future
    	release of ALL-IN-1 and the FCS..."
    
    Graham

    My customer has the same problem. Using MS STA or 
    
    fcs :== "$oafc$server"
    $fcs oa$data:node$server73.dat 
    
    produces the error
    
    23-AUG-1993 15:17:24.67  Server: KERMIT::"73="  Error: %OAFC-E-INTERR,
    Internal
    error in File Cabinet Server  Message: PCAdd; Partition already in name
    cache. %OAFC-E-INTERR, Internal error in File Cabinet Server
    
    I used MS D to delete the server, then 
    <do OA$LIB:OAFC$SERVER_CREATE_73.SCP to recreate the data file but the same
    happens and the server remains STOPPED.
    There are also a large number of these errors :-
    
    %MCC-E-FATAL_FW, fatal framework condition:  !AS  Message: SrvTimeoutSysMan;
    receive alert to terminate thread
    
    They are running ALL-IN-1 British 3.0-1 on VMS 5.5-2 in a cluster where
    SYS$CLUSTER_NODE="KERMIT::"
    
    $SHOW CLUST gives
    
    View of Cluster from system ID 6147  node: KERMIT          23-AUG-1993
    17:23:32
    
         SYSTEMS       MEMBERS 
    
      NODE   SOFTWARE   STATUS 
    
     KERMIT  VMS V5.5  MEMBER  
     SYSDSK  RFX T387          
     SAS1    RFX T387          
     SAS2    RFX T387          
     SAS3    RFX T387          
     OA1     RFX T387          
     OA2     RFX T387          
     OA3     RFX T387          
     ZOOT    VMS V5.5  MEMBER  
     ROWLF   VMS V5.5  MEMBER  
     BEAKER  VMS V5.5  MEMBER  
     HARRY   VMS V5.5  MEMBER  
     BUNSEN  VMS V5.5  MEMBER  
     ANIMAL  VMS V5.5  MEMBER  
     FOZZIE  VMS V5.5  MEMBER  
     DEADLY  VMS V5.5  MEMBER  
    
    
    SYSGEN>  SHO SCSNODE
    Parameter Name            Current    Default     Min.     Max.     Unit Dynamic
    --------------            -------    -------    -------  -------   ---- -------
    SCSNODE                 "KERMIT  "    "    "    "    "    "ZZZZ" Ascii
    
    SYSGEN>  SHO SCSSYSTEMID
    Parameter Name            Current    Default     Min.     Max.     Unit Dynamic
    --------------            -------    -------    -------  -------   ---- -------
    SCSSYSTEMID                  6147          0        -1       -1 Pure-numbe
    
    
    SYSGEN>  SHO VAXCLUSTER
    Parameter Name            Current    Default     Min.     Max.     Unit 
    Dynamic
    --------------            -------    -------    -------  -------   ---- 
    -------
    VAXCLUSTER                      2          1         0         2
    Coded-valu
    
    I changed the entry in profile.fdl to be owner ALLIN1 (from SYSTEM) but
    that shouldn't/didn't help.
    
    
    File cab server id is :-
    
                    image name: "OAFC$SERVER"
                    image file identification: "OAFC V1.0-2"
    
    
    What else do I need to check please ?
    
    Thanks,
    
    Clive
    (I've checked note 730 too)

    I find out today that my colleague in Messaging had to redefine
    SYS$CLUSTER_NODE on the node KERMIT.
    Message Router and ALL-IN-1 run on a single node called KERMIT in the
    MUPPET (3 node) cluster. On the node KERMIT they redefined
    SYS$CLUSTER_NODE from MUPPET:: to equal KERMIT:: which apparently fixes 
    an MR problem. It may be the cause of our problem? 
    
    Is it OK to give this logical a blank value on KERMIT ?  
    
    Thanks,
    
    Clive

    
    This is the problem, when the server starts up, it loops to get all the
    node names in teh cluster, then gets the cluster alias, if it exists.
    
    By having a node the same as the alias, the FCS tries to put it in
    twice.
    
    >    Is it OK to give this logical a blank value on KERMIT ?  
    
    Yup, there could be problems. If uses connect via alias, no drawers
    are named via alias, or you broker out/in to/from this node, there
    will be a problem. In short, i wouldn't recommend it.
    
    --Bob