Re: BUG #16643: PG13 - Logical replication - initial startup never finishes and gets stuck in startup loop

I've made an important observation!

Since I had the impression this setup was already working with RC1 of PG
13, I re-installed RC1 and did the same test. And it's working fine!

So it seems that something has changed from RC1 to the final version that
breaks logical replication in some cases.

Best,
Henry

Am Mi., 30. Sept. 2020 um 08:34 Uhr schrieb Henry Hinze <
henry(dot)hinze(at)gmail(dot)com>:

> Hi Tom,
>
> thanks for your reply!
>
> For my comparison test of PG 12 and 13 I did not change any default
> setting except wal_level.
>
> But I also did some tests and increased wal_keep_size
> and max_slot_wal_keep_size to 1GB. And I set wal_sender_timeout to 1h but
> without success. The setup works in PG 13 only with a small amount of data.
>
> Best,
> Henry
>
> Am Di., 29. Sept. 2020 um 22:09 Uhr schrieb Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
>
>> PG Bug reporting form <noreply(at)postgresql(dot)org> writes:
>> > The following setup of logical replication for one table works nicely
>> with
>> > PG 12.4. In PG 13.0 the initial startup never finishes. Instead it
>> starts
>> > over and over again.
>>
>> It kinda looks like something is timing out before it's managed to
>> catch up with the (rather large) initial state of table t. What
>> non-default replication settings are you using?
>>
>> regards, tom lane
>>
>
>
>