Conference iosg::all-in-1_v30

    POSTMASTER_MAIL_NBS_FAIL contains mail that the ALL-IN-1 Fetcher failed
    to deliver. The Fetcher will re-try these failed messages (because they
    are also left on the fetcher queue in PENDING.DAT) and may subsequently
    succeed to deliver them or may reach the re-try limit and return a
    non-delivery to the sender.
    
    It may be that the fetcher is only delivering to some recipiants, all
    recipiants or none.
    
    If you are confident that these messages were fully delivered then they
    can be deleted. It may be worth checking OA$MTI_ERR for errors at
    about the time the messages in this folder were created just in case
    you have a problem on the system that is causing mail to fail to
    deliver.
    
    Basically this folder contains a list of mail that the fetcher had
    problems with and needs checking out to see if it was delivered and if
    not why not.
    
    HTH
    
    Terry

     In the OA$MTI_ERR, we found a lot of error messages.

27-MAY-1992 19:01:08       %OA-E-IOCOPY, Error copying to file ""
27-MAY-1992 19:01:08       -RMS-F-DEV, error in device name or inappropriate 
                            device type for operation

     All disk drives have not been dismounted. Logical OA$SHAR* is pointed to 
existing directory. What does ALL-IN-1 copy? How can I trace the problem?

Thanks,
  Steven Lau

I think the only time a copy is done by the Fetcher is when the arriving 
message is being shared. So those error messages hint, to me, that at the time 
the Fetcher ran, that the shared area it was trying to write to didn't exist. 
It may well have been fixed since by a reboot (and hence rerun of 
A12V*START.COM) so you won't be able to find it now.

If that did happen, the message would be retained on the queue in the Pending 
file and (almost certainly) delivered on a subsequent run of the fetcher.

grahame

     In the customer's system, the shared directories are ran from OA$SHARE1 to
OA$SHARE84. But in between, the logical OA$SHARE21 has not defined, and that 
directory does not exist. Would that cause the problem? If yes, how can I add
the directory back?

Thanks,
   Steven Lau

    The first step would be to create the shared directory on the device
    where you wish it to be placed. Normally this would be under the top
    level OA$SHARE directory while logged into the ALLIN1 account.
    
    $Set def disk:[OA$SHARE]
    $Create/dir [.share21]
    
    Then enter the ALL-IN-1 menus and access the shared directory master
    file entry form to add the directory.
    
    SM_SHARED_DIR_MASTER <RETURN>
    On the "Shared directory logical:" field enter OA$SHARE21 <RETURN>
    
    The rest of the fields should remain blank at this point and you will
    be given the options to Add, Change, COpy, Delete, or Inquire:
    
    Choose "A" to add and enter the location of the new shared directory
    area on the "Logical value:" field. Do not attempt to enter any other
    information in the other fields, just <RETURN> out to the blank form
    again. After you <EXIT SCREEN> back to the menu you can run the
    folowing script on each cluster node (running ALL-IN-1) to define the
    missing logical.
    
    <DO OA$SM_DEFINE_SHARED_LOGICALS.SCP
    
    A missing shared directory will definitely cause problems in any area
    of mail if it is located in the current write window on the system.
    
    Hope this helps.