Conference iosg::all-in-1_v30

2846.0. "ALL-IN-1 Index Server Internal Error" by TINNIE::SETHI (Ahhhh (-: an upside down smile from OZ) Fri Jun 11 1993 01:56

    Hi FCS people,                   
    
    A customer has a number of errors reported in the OAFC$SERVER_ERROR.LOG
    logfile.  Since this should not happen too often I am posting the
    errors here.  The customer has ALL-IN-1 IOS 3.0-1 installed.
    
    ALL-IN-1 Index Server Internal Error:
    Error locking DAB during cache garbage collection:
    The lock on the following drawer has become invalidated by another
    process.  Note that the lock has been granted and OafcNormal
    will be returned to the client, however, all other processes wishing to
    share this lock will also be granted invalid locks until all processes
    sharing this lock are terminated.
    Drawer directory:
    DISK$USER12:[USER_ALLIN1.WOP202AI1.A1.ZUFXCNHV6]0 Drawer owner: WOP202AI1
    
    The above message appears for the same user but the Drawer directory
    changes.
    
    ALL-IN-1 Index Server Internal Error:
        Error locking DAB during cache garbage collection:
    
    The customer has a cluster with a 100 or more users per node and has
    seen hundreds of %MCC-E-FATAL_FW as reported in note 2563, ever since
    they upgraded to 3.0-1.  The FCS has had to be shutdown and restarted
    a number of times as shared file cabs are no longer shared. By the way
    the OAFC$SERVER..LOG has grown to 15,000 blocks and I have asked the
    customer to rollover the logfile.
    
    What does this all mean ?  Does it mean we have a problem with 3.0-1 ?
    Is there anything you would like me to do to track down the problem ?
    
    Regards,
    
    Sunil

    
    re the message in oafc$server_error.log
    
    I have gotten this message several times, it is PROBABLY nothing to
    worry about. I have checked the code and here is what the comments
    say around that code.
    
    this is caused by a user and IOS locking the drawer (ie not using the
    FCS for the lock), then terminating the A1 session abnormally, 
    for instance cntrl-Y out,  A1 exception handlers have somehwo 
    managed to keep the image "semi-alive", thus invalidating the lock
    block.
    
    Note that nothing bad will happen with this, it is simply that the
    two parts that are cooperating with the lock manager, one of them
    is not quite cooperating correctly. The chances are the IOS really
    has stopped, so accessing the drawer is fine.
    
    The best way to clear this is to re-start ALL-IN-1. Also note that
    according to the code, the drawer in question is the main drawer, it
    is the only one that IOS will lock on its own. Therefore, if you are
    seeing this while accessing another drawer, speak up now becuase that
    means the above is not true.
    
    --Bob