Re: controlling snapshots

Hi Roman,

Did you verify your user_jobs dictionary view? Snapshot refreshes are being processed as pl/sql jobs. If a pl/sql job fails it will be automatically resubmitted 16! times before the job is considered broken. The user_jobs view also has an error code if the job is broken. You could consider creating a snapshot refresh group. In this case the refreshes will be executed sequentially and they won't block each other, or saturate the network.

Hth,

Sybrand Bakker, Oracle DBA

Roman wrote:

> We have following situation:
> There are a couple of snapshots that are to be replicated every day from a
> central database to many decentral databases. There is a deadpoint where all
> replication should be completed.
>
> Problem:
> Because of heavy network traffic, some snapshots don't succeed in time.
>
> Aim:
> An automatic DB-procedure which controlls the execution of the snapshots and
> reexecutes them if not ok.
>
> How should it be done?
>
> Thanks for any help,
> Roman

Received on Fri Jan 08 1999 - 13:54:38 CST