On Mon, May 15, 2017 at 5:04 PM, Masahiko Sawada <sawada.mshk@gmail.com> wrote:
> On Mon, May 15, 2017 at 8:22 PM, Kuntal Ghosh
> <kuntalghosh.2007@gmail.com> wrote:
>> On Mon, May 15, 2017 at 4:39 PM, Masahiko Sawada <sawada.mshk@gmail.com> wrote:
>>
>> While testing with logical replication, I've found that the server
>> hangs if we create a subscription to the same server. For example,
>>
>> $ ./pg_ctl -D Test start -o "-p 5432"
>> $ ./psql -p 5432 postgres -c "CREATE PUBLICATION abc for all tables
>> with (publish='delete');"
>> $ ./psql -p 5432 postgres -c "create subscription sub connection
>> 'dbname=postgres port=5432 user=edb password=edb' publication abc with
>> (slot_name='abcslot');"
>> NOTICE: synchronized table states
>> LOG: logical decoding found initial starting point at 0/162DF18
>> DETAIL: Waiting for transactions (approximately 1) older than 560 to end.
>>
>> And, it hangs. Is this an expected behavior?
>
> I guess this is what we're discussing on [1].
> Petr answered on that,
> ===
> Yes that's result of how logical replication slots work, the transaction
> that needs to finish is your transaction. It can be worked around by
> creating the slot manually via the SQL interface for example and create
> the subscription using WITH (NOCREATE SLOT, SLOT NAME = 'your slot') .
> ===
>
> [1] https://www.postgresql.org/message-id/20170426165954.GK14000%40momjian.us
>
Ohh I see. Thank you.
--
Thanks & Regards,
Kuntal Ghosh
EnterpriseDB: http://www.enterprisedb.com
