From: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> |
---|---|
To: | Robert Haas <robertmhaas(at)gmail(dot)com> |
Cc: | Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>, Andrey Borodin <x4mmm(at)yandex-team(dot)ru>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Logical WAL sender unresponsive during decoding commit |
Date: | 2022-10-20 05:37:24 |
Message-ID: | CAA4eK1+B6qECQi=pURJV6w3wV+ZGS6C78-NiDi5zUhUJpvnaOg@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, Oct 20, 2022 at 5:17 AM Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> >
> > Pushed.
>
> I think this was a good change, but there's at least one other problem
> here: within ReorderBufferRestoreChanges, the while (restored <
> max_changes_in_memory && *segno <= last_segno) doesn't seem to contain
> a CFI. Note that this can loop either by repeatedly failing to open a
> file, or by repeatedly reading from a file and passing the data read
> to ReorderBufferRestoreChange. So I think there should just be a CFI
> at the top of this loop to make sure both cases are covered.
>
Agreed. The failures due to file operations can make this loop
unpredictable in terms of time, so it is a good idea to have CFI at
the top of this loop.
I can take care of this unless there are any objections or you want to
do it. We have backpatched the previous similar change, so I think we
should backpatch this as well. What do you think?
--
With Regards,
Amit Kapila.
From | Date | Subject | |
---|---|---|---|
Next Message | Fabrice Chapuis | 2022-10-20 05:46:50 | Re: Logical replication timeout problem |
Previous Message | Thomas Munro | 2022-10-20 04:54:47 | Re: Understanding, testing and improving our Windows filesystem code |