The amount of space needed in the recovery log depends on the number of
transactions in flight; the size of the files being backed up really isn't
relevant.

I didn't know it was possible to run effectively with a 12 MB recovery log!
On an AIX ADSM server, we have always used at least 200 MB for the recovery
log, and now that we're using rollforward, the recovery log is up to 2 GB!

(However, we also use 70 GB DLT tapes, which means that a reclaim can
involve MANY MANY transactions, and can be running at the same time as
backups, backupstgpool, and migration, resulting in a LOT of concurrent data
base transactions.)

But Yes, I think your problem can easily be cured by increasing the size of
your recovery log.

> -----Original Message-----
> From: Thomas Denier [SMTP:Thomas.Denier AT MAIL.TJU DOT EDU]
> Sent: Thursday, May 20, 1999 11:33 AM
> To:   ADSM-L AT VM.MARIST DOT EDU
> Subject:      Recovery log contention
>
> My site has a 3.1.2.20 ADSM server under MVS. We have a 12 megabyte
> recovery
> log. We do not use roll-forward mode. Until last weekend, I never saw the
> recovery log utilization go above 40 percent. Last weekend we ran a backup
> of
> a mail repository with about a million files, with most of the files
> having
> sizes of a couple of kilobytes. The backup was still running when the
> 'backup
> stgpool' commands used to create offsite backup tapes started. The
> recovery
> log utilization increased rapidly until the recovery log was full. At that
> point the client backup ended and the recovery log occupancy dropped off.
> The
> log on the client system reported that the backup was terminated because
> it
> detected a system error.
>
> This weekend our database administrators are planning a major upgrade and
> reorganization of an Oracle server, and plan to run ADSM backups at
> various
> points in the process. The times of these backups will depend on how long
> the
> various steps in the process take, and are not known in advance. It is
> entirely possible that some of them will overlap the storage pool backups.
> The
> backups will contain a couple of hundred files with sizes ranging from a
> few
> megabytes to about two gigabytes. With this kind of size distribution, are
> we
> likely to see problems with recovery log utilization if the Oracle server
> backups overlap with storage pool backups? Can the problem be cured by
> increasing the recovery log size to some reasonable value?