| From: | Nitin Jadhav <nitinjadhavpostgres(at)gmail(dot)com> |
|---|---|
| To: | Heikki Linnakangas <hlinnaka(at)iki(dot)fi> |
| Cc: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: [PATCH] Fix recovery connection detail in EXEC_BACKEND builds |
| Date: | 2026-10-15 12:47:54 |
| Message-ID: | CAMm1aWY3aCaB78_B4K6uZTc=L4WeaQgk2YZ78K8Rfg+UFTFfSA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Heikki,
Thanks for reviewing the patch.
I had considered adding a separate CAC_* code as well, but since the
existing code had already combined these cases under
CAC_NOTHOTSTANDBY, I initially chose the smaller change and preserved
that structure.
I agree that distinguishing the reasons in CAC_state is cleaner. I
will restore CAC_NOTCONSISTENT and use CAC_NOTHOTSTANDBY specifically
for the case where hot standby is enabled and consistency has been
reached, but the recovery snapshot is not yet ready. This would also
avoid passing reachedConsistency through BackendStartupData.
I will revise the patch accordingly and send a v2 soon.
Best Regards,
Nitin Jadhav
Microsoft
| From | Date | Subject | |
|---|---|---|---|
| Next Message | David Rowley | 2026-10-15 12:49:43 | Re: Avoid repeated scans of the output buffer in oidvectortypes() |
| Previous Message | Chao Li | 2026-10-15 12:15:31 | Re: Avoid repeated scans of the output buffer in oidvectortypes() |