Re: Bug? pg_rewind produces unusable but starting database with standby recovery

Hi all,

On Tue, Sep 8, 2026 at 5:21 PM Zsolt Parragi <zsolt(dot)parragi(at)percona(dot)com> wrote:
>
> Thank you both for the review! After reviewing the issue again, I
> agree that my 0002 was wrong.
>
> 0001 received a trivial rebase as there were no conflicts, and I
> simply removed 0002 in favor of ChangAo Chen's thread.

Thank you for the updated patch.
I reviewed and tested the patch. The change in StartupXLOG() looks
correct. The minRecoveryPoint check is now enforced independently of
ArchiveRecoveryRequested, which covers the pg_rewind case where
minRecoveryPoint is set without archive-recovery signal files. The
existing online-backup WAL handling is also preserved. I also verified
the new regression test for the missing source WAL scenario. The test
covers both failure when the required WAL is unavailable and also the
successful recovery when the WAL is available through archive
recovery. Overall, the patch looks correct to me and I did not find
any issues during review and testing.

Regards
Solai