On Wed, Dec 13, 2023 at 5:39 AM Jakub Wartak
<jakub.wartak@enterprisedb.com> wrote:
> > I can't exactly say that such a hint would be inaccurate, but I think
> > the impulse to add it here is misguided. One of my design goals for
> > this system is to make it so that you never have to take a new
> > incremental backup "just because,"
>
> Did you mean take a new full backup here?
Yes, apologies for the typo.
> > not even in case of an intervening
> > timeline switch. So, all of the errors in this function are warning
> > you that you've done something that you really should not have done.
> > In this particular case, you've either (1) manually removed the
> > timeline history file, and not just any timeline history file but the
> > one for a timeline for a backup that you still intend to use as the
> > basis for taking an incremental backup or (2) tried to use a full
> > backup taken from one server as the basis for an incremental backup on
> > a completely different server that happens to share the same system
> > identifier, e.g. because you promoted two standbys derived from the
> > same original primary and then tried to use a full backup taken on one
> > as the basis for an incremental backup taken on the other.
> >
>
> Okay, but please consider two other possibilities:
>
> (3) I had a corrupted DB where I've fixed it by running pg_resetwal
> and some cronjob just a day later attempted to take incremental and
> failed with that error.
>
> (4) I had pg_upgraded (which calls pg_resetwal on fresh initdb
> directory) the DB where I had cronjob that just failed with this error
>
> I bet that (4) is going to happen more often than (1), (2) , which
> might trigger users to complain on forums, support tickets.
Hmm. In case (4), I was thinking that you'd get a complaint about the
database system identifier not matching. I'm not actually sure that's
what would happen, though, now that you mention it.
In case (3), I think you would get an error about missing WAL summary files.
> > Huzzah, the cfbot likes the patch set now. Here's a new version with
> > the promised fix for your non-reproducible issue. Let's see whether
> > you and cfbot still like this version.
>
> LGTM, all quick tests work from my end too. BTW: I have also scheduled
> the long/large pgbench -s 14000 (~200GB?) - multiple day incremental
> test. I'll let you know how it went.
Awesome, thank you so much.
--
Robert Haas
EDB: http://www.enterprisedb.com