Greetings,
* Laurenz Albe (laurenz.albe@cybertec.at) wrote:
> I too only learned about this recently, while the problem with exclusive
> backups has been known at least since 2008 (c979a1fe), and nobody felt
> this to be a terrible problem back then.
My recollection was that back then it wasn't clear what to *do* about
the problem. Once we had a solution (non-exclusive backups) we
deprecated the exclusive backup mechanism.
> I believe that the danger is greatly overrated. It is not like you end
> up with a corrupted database after a crash, and you get a pretty helpful
> error message. Many people are happy enough to live with that.
Unfortunately, people get corrupted databases all the time because the
backups are corrupted. The problem with databases not restarting
correctly is another issue with the exclusive backup method but it's
certainly not the only one.
> I'm on board with deprecating and removing it eventually, but I see no
> problem in waiting for the customary 5 years.
We don't actually have a 'customary 5 years' rule of any sort.
> And yes, a prominent warning in the next major release notes would be
> a good thing.
This is two years too late, at best. That is, maybe it would have made
sense to highlight that the exclusive backup method was deprecated when
it was made so, but that ship has sailed. Further, we don't put out
announcements saying "we're going to remove X in the next release" and I
don't see it making sense to start now.
Thanks!
Stephen
