Jerry,

Try searching IBMLINK, if you have access.
I had this problem, too, and my server is AIX 2.1.5.12.
No matter how many incrementals you run, it just keeps kicking off
incrementals.

I believe I found a hit in IBMLINK that said that a long-running
transaction can keep entries "pinned" in the recovery log.

So no matter how many incrementals you do, it doesn't clear the recovery
log.
You can kill the problem by bouncing the server, or killing the
long-running process (in your case, possibly the Expire Inventory).

After it happened a couple of times, I just made my recovery log a lot
bigger, and haven't had the problem since


 =======================================================================
Wanda Prather
Johns Hopkins Applied Physics Lab
301-953-6000 X8769
wanda_prather AT jhuapl DOT edu

"Intelligence has much less practical application than you'd think."
              - Scott Adams/Dilbert
 =======================================================================
.


        ----------
        From:   Jerry Lawson[SMTP:jlawson AT thehartford DOT com]
        Sent:   Saturday, November 15, 1997 10:51 PM
        To:     ADSM-L AT vm.marist DOT edu
        Subject:        Recovery Log anomaly

        Date:     November 15, 1997             Time: 9:44 PM
        From:     Jerry Lawson
                  The Hartford Insurance Group
        (860)  547-2960          jlawson AT thehartford DOT com

------------------------------------------------------------------------
-----
-----
        Just found a strange happening with my Recovery log.....
        Just found a strange happening with my Recovery log.....

        We have a 1500MB Recovery Log, which is Mirrored.  We run Roll
Forward mode.
        My server is MVS V2 at L13.  The normal processing we do fills
this
        approximately 70% in a day, we have the trigger set at 80%.

        At 09:00 today (Saturday), a scheduled full DB backup started.
It completed
        approximately 1 ? hours later.  At 11:00, we started an Expire
Inventory.

        Approximately 12:30, an incremental DB backup was triggered
because we were
        at the threshold (80%) ?????

        It backed up about 56K pages (there were 2.1M pages backed up on
the full).
        The messages indicated a normal conclusion of the incremental
(the full ended
        normally too).

        8 minutes later, another incremental kicked off, and backed up
4800 pages.
        It completed normally, and another incremental kicked off - this
one backed
        up 1600 pages!

        This continued until 10 incrementals had kicked off,   at which
point a full
        backup was initiated.  This ran for about 2 hours, and then
finished.  There
        was a message indicating that a recovery from a full recovery
log was
        completed.  Then another incremental was kicked off 20 minutes
later.  ?????

        There is now a 4 hour gap in the activity log, but something was
going on, as
        when the recovery log picks up 4 hours later, apparently when I
found a
        problem and started an incremetal that backed up 69 pages.
There was
        activity in the 4 hours though, as the process numbers jumped by
about 60
        (remember that the expire inventory was running, and
reclaimations were going
        on.

        When I came in, the crrent and max utilization of the recovery
log was at
        100%.

        Fortunately, ADSM was still up and seemed to be otherwise ok.  I
then went
        and scrounged another 280MB, and added a recovery log DB.  As
soon as the
        process was completed, utilization was now at 2.2% ot 1800+ MB.

        It would appear that the Recovery Log was not reset when the
backup was
        completed.  Has anyone else seen something like this?




------------------------------------------------------------------------
-----
-----
                                                             Jerry
                                                             Jerry