Re: Ignore heap rewrites for materialized views in logical replication

On Tue, May 31, 2022 at 8:28 PM Euler Taveira <euler(at)eulerto(dot)com> wrote:
>
> On Tue, May 31, 2022, at 11:13 AM, Amit Kapila wrote:
>
> I think we don't need the retry logical to check error, a simple
> wait_for_caught_up should be sufficient as we are doing in other
> tests. See attached. I have slightly modified the commit message as
> well. Kindly let me know what you think?
>
> Your modification will hang until the test timeout without the patch. That's
> why I avoided to use wait_for_caught_up and used a loop for fast exit on success
> or failure.
>

Right, but that is true for other tests as well and we are not
expecting to face this/other errors. I think keeping it simple and
similar to other tests seems enough for this case.

> I'm fine with a simple test case like you proposed.
>

Thanks, I'll push this in a day or two unless I see any other
suggestions/comments. Note to others: this is v10 fix only. As
mentioned by Euler in his initial email, this is not required from v11
onwards due to a different solution for this problem via commit
325f2ec555.

--
With Regards,
Amit Kapila.