> On 31 Aug 2016, at 03:28, Craig Ringer <craig@2ndquadrant.com> wrote:
>
> On 25 Aug. 2016 20:03, "Stas Kelvich" <s.kelvich@postgrespro.ru> wrote:
> >
> > Thanks for clarification about how restart_lsn is working.
> >
> > Digging slightly deeper into this topic revealed that problem was in two phase decoding, not it logical decoding
itself.
>
> Good to know. The behavior didn't make much sense for logical decoding but bugs usually don't.
>
> Do you plan to submit a patch for logical decoding of prepared transactions to 10.0?
>
I do, probably on CF2. There is issue with locks that Andres pointed me to; also twophase.c was written before logical
decoding happened and it deserves some refactoring to avoid duplicated functionality between 2pc decoding and restoring
of
old prepared xact — I want to include that in patch too.
--
Stas Kelvich
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company
