On Mon, Aug 28, 2017 at 8:33 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Michael Paquier <michael.paquier@gmail.com> writes:
>> Attached is a patch to make this code path wait that the transaction
>> has been replayed. We could use as well synchronous_commit = apply,
>> but I prefer the solution of this patch with a wait query.
>
> Petr proposed a different patch to fix the same problem at
> https://www.postgresql.org/message-id/1636c52e-c144-993a-6665-9358f322deda%402ndquadrant.com
>
> Which one is better?
Petr's patch is here to fix the race condition in test 006 for logical
decoding. My patch involves the failures that dangomushi sees since
yesterday with 2PC test 009. (Note: It took me a while to put this
animal in an environment with a more stable connection, so now I hope
that we'll get again daily reports)
--
Michael