| From: | Julien Rouhaud <rjuju123(at)gmail(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: Isn't wait_for_catchup slightly broken? |
| Date: | 2022-01-11 06:25:02 |
| Message-ID: | 20220111062502.obh4gegzcaova5dk@jrouhaud |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, Jan 10, 2022 at 02:31:38PM -0500, Tom Lane wrote:
>
> So I think we need to fix it to capture the target WAL position
> at the start, as I've done in the attached patch.
+1, it looks sensible to me.
> In principle
> this might make things a bit slower because of the extra
> transaction required, but I don't notice any above-the-noise
> difference on my own workstation.
I'm wondering if the environments where this extra transaction could make
a noticeable difference are also environments where doing that extra
transaction can save some iteration(s), which would be at least as costly.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Gurjeet Singh | 2022-01-11 07:07:22 | Patch: Code comments: why some text-handling functions are leakproof |
| Previous Message | Amit Kapila | 2022-01-11 06:12:01 | Re: Skipping logical replication transactions on subscriber side |