Conference clusta::acms

4123.0. "QTI file growth" by HSIC02::BOWYER () Thu Mar 13 1997 07:20

    My customer (a bank with a 7 x 24 operation) uses the QTI quite a lot. 
    There are about 1500 entries queued per day, all of which should be
    processed within a few minutes.  In the past they found the data file
    for the queue (i.e. NOT the QDF-file, but the file the QDF points to)
    grew pretty big (say 30000 blocks) and got slower and slower.
    
    As a workaround we used a CONVERT/RECLAIM on the rare occasions when
    the whole cluster was down.  On one occasion we hit a bug in
    CONVERT/RECLAIM and SSO advied us to use CONVERT instead of
    CONVERT/RECLAIM.  However, the customer didn't want to do that. 
    Moreover they've bought more nodes so that the cluster "never" goes
    down.
    
    This morning I was looking at the file again and found it now no longer
    seems to grow.  They re-created it duing a software upgrade two weeks
    ago and it's still at only 1500 blocks.  Has the QTI changed somewhere
    along the line (we now have ACMS V4.1, the problem was first noticed in
    3.3) so that the problem has gone away?  Before you ask, they don't use
    RMS journalling and there's no need to back up the queue file.
    
    I'd like to be able to tell them the queue file will behave itself
    without any special actions.  Can I do that?