My experience is with ADSM V2 on AIX, but it may be helpful.
I strongly recommend the use of ROLLFORWARD mode because it helps you so
much if you need to do a data base recovery.
However, we also had a few problems when we first turned on ROLLFORWARD,
until we got things sized right.
The first thing I would do is check is to see what was happening on the ADSM
server over the weekend when you got the extra DB backups triggered.
When there are very long running operations in progress, such as a MOVE DATA
or a tape RECLAIM, it appears that transactions associated with those long
running operations do NOT all get cleared from the log when a data base
backup is done.
We found that a data base backup done while a tape reclaim was in progress
did not free up all log space; the result was that the DB backup trigger
caused a second and then third DB backup to be taken one right after the
other.
All that is necessary to correct this is to make the log big enough that you
don't have the problem anymore.
In your case, it certainly appears your log is big enough at 3.5 GB. But
with your DB BACKUP trigger set to only 20%, you can only use 700 MB of that
space before a DB backup is triggered.
In your case I would suggest RAISING your DB backup trigger, to maybe 40%.
Then you would have 1.4 GB of log space to fill before triggering a DB
backup.
If that doesn't change things, or if the problem recurs when there appear to
be NO long-running processes on the server, I would report it to IBM as a
bug.
Hope this helps, maybe someone else will have some other suggestions...
***************************************************************
Wanda Prather
The Johns Hopkins Applied Physics Lab
443-778-8769
wanda_prather AT jhuapl DOT edu
"Intelligence has much less practical application than you'd think" -
Scott Adams/Dilbert
***************************************************************
> -----Original Message-----
> From: Moodley Desigan * Datavia [SMTP:DesiganM AT transnet.co DOT za]
> Sent: Monday, November 30, 1998 3:42 AM
> To: ADSM-L AT vm.marist DOT edu
> Subject: Roll forward recovery
>
> Hello list members,
>
> We're running ADSM server version 3.1.2.0 on MVS OS 390. We recently
> changed
> our recovery log mode from normal to roll forward. We set a database
> backup
> trigger of 20%. When the log reached 20%, a full database backup is
> triggered. This usually happens about every 24 hours. After the backup,
> the
> recovery log percentage dropped to about 1%. This continued for about 4
> days. However, during the weekend, the recovery log percentage did not
> drop
> after a database backup causing database backups to be triggered
> immediately
> after a successful backup. Here is an extract of the job log:
>
> ANR4550I Full database backup (process 530) complete, 2666266 pages
> copied.
>
> ANR0985I Process 530 for DATABASE BACKUP running in the BACKGROUND
> completed
>
> ANR0985I with completion state SUCCESS at 07:51:23.
>
> ANR4556W Warning: the database backup operation did not free sufficient
>
> ANR4556W recovery log space to lower utilization below the database backup
>
> ANR4556W trigger. The recovery log size may need to be increased.
>
> We were eventually forced to bounce ADSM to get the recovery log back down
> to 0.1 %. Here is the recovery log status:
>
> Available Space (MB): 3,520
> Assigned Capacity (MB): 3,520
> Maximum Extension (MB): 0
> Maximum Reduction (MB): 3,484
> Page Size (bytes): 4,096
> Total Usable Pages: 900,608
> Used Pages: 8,420
> Pct Util: 0.9
> Max. Pct Util: 0.9
> Physical Volumes: 4
> Log Pool Pages: 512
> Log Pool Pct. Util: 0.11
> Log Pool Pct. Wait: 0.00
> Cumulative Consumption (MB): 125,835.95
> Consumption Reset Date/Time: 04/16/1998 09:07:19
>
> Here is our database status:
>
> Available Space (MB): 16,044
> Assigned Capacity (MB): 15,004
> Maximum Extension (MB): 1,040
> Maximum Reduction (MB): 4,588
> Page Size (bytes): 4,096
> Total Usable Pages: 3,841,024
> Used Pages: 2,666,346
> Pct Util: 69.4
> Max. Pct Util: 69.4
> Physical Volumes: 11
> Buffer Pool Pages: 8,192
> Total Buffer Requests: 10,998,632
> Cache Hit Pct.: 99.15
> Cache Wait Pct.: 0.00
> Backup in Progress?: No
> Type of Backup In Progress:
> Incrementals Since Last Full: 2
> Changed Since Last Backup (MB): 39.00
> Percentage Changed: 0.37
> Last Complete Backup Date/Time: 11/30/1998 08:34:27
>
> We are still running in roll forward mode. Can someone advise us as to
> whether we should keep going in roll forward mode or revert to normal
> mode.
> Also any advise as to how to manage roll forward mode will be appreciated.
>
> Thanks.
>
> Kind regards,
> Desigan Moodley
> * <mailto:DesiganM AT Transnet.co DOT za> DesiganM AT Transnet.co DOT za
|
