"Jim Starkey" <jas@...> wrote:

>
> I would like to see a careful, thoughtful design, something that would
> allow any backup that ran to completion to be restored. This will
> certainly require the ability to disable various constraints when
> necessary. It certainly would require the ability to disable foreign
> key checking on specific foreign keys (a very good thing in its own
> right).

Agreed 100%.

> It is important to metadata updates performed during DDL statements be
> completely validated, but this shouldn't apply to direct system table
> updates.

Then we need to support every metadata update via DDL. As you know, this is
not the case currently.

> I would also like to make
> explicitly illegal any reference to tables within a transaction in which
> the table's system tables have been modified.

The multi-metadata changes are already broken in Vulcan. Let's discuss this
before going further. I'll post about the known issues tomorrow.

> And make all DDL
> statements effectively auto-commit.

We've already been there before. But I agree to get back to this question
again - it deserves more discussion.

> But to answer your question, if you backup a broken procedure, after a
> restore, you still have a broken procedure.

Currently, you get a partially restored database without any procedures.

> The bottom line is that I consider unrestorable backups to be a
> near-criminal offense.

Everyone would second this. But our practice is that we never had reliable
backups. Maybe during the past 10 years, maybe more...


Dmitry