On Wed, 31 Mar 2004, Philip Warner wrote:
> At 12:13 AM 31/03/2004, Bruce Momjian wrote:
> >Yes, they have to check for a proper exit from pg_dump, but there is
> >still a file sitting around after the dump, with no way to tell if it is
> >accurate.
>
> Why don't we write a hash into the header or footer. Then use something like:
>
> pg_restore --verify <dump-file>
>
> if file integrity is the objective.
I like this idea. Nice to be able to check the md5 sig to make sure a
backup is not corrupted OR short.