Re: Would-be standby-node phoning home every 5 seconds. Why?

Yes, it's correct! Glad to hear it's fixed. Have a nice day!

Ron Johnson <ronljohnsonjr(at)gmail(dot)com>, 19 Ara 2023 Sal, 19:28 tarihinde
şunu yazdı:

> Errors disappeared after -- on node b -- I deleted $PGDATA/standby.signal
> (makes sense) and primary_conninfo from postgresql.auto.conf, then
> restarted the instance.
>
> On Tue, Dec 19, 2023 at 11:14 AM Ron Johnson <ronljohnsonjr(at)gmail(dot)com>
> wrote:
>
>> wal_level = minimal.
>>
>> On Tue, Dec 19, 2023 at 11:09 AM SOzcn <selahattinozcnma(at)gmail(dot)com>
>> wrote:
>>
>>> Hmm, It seems like NodeB is attempting to initialize replication despite
>>> the termination and dropping of the replication slot. The error message
>>> suggests that it's trying to establish a standby connection but
>>> encountering a limitation regarding max_wal_senders, which is currently set
>>> to 0.
>>>
>>> Could you restart your NodeB to check this? If It does not work there
>>> are also some connection strings from applications.
>>>
>>> Also is your wal_level is replica in the Primary?
>>>
>>> Ron Johnson <ronljohnsonjr(at)gmail(dot)com>, 19 Ara 2023 Sal, 18:26 tarihinde
>>> şunu yazdı:
>>>
>>>> max_wal_senders is currently 0. That's why it's complaining.
>>>>
>>>> On Tue, Dec 19, 2023 at 9:57 AM SOzcn <selahattinozcnma(at)gmail(dot)com>
>>>> wrote:
>>>>
>>>>> Hello Ron, Can you check your "max_wal_senders" parameter value?
>>>>> This is must come from the "max_wal_senders" parameter, if you set
>>>>> this as 5 seconds it's going to check this for each 5 second.
>>>>>
>>>>> If you want to disable this, parameter value must be 0.
>>>>>
>>>>> Have a nice day!
>>>>>
>>>>> Ron Johnson <ronljohnsonjr(at)gmail(dot)com>, 19 Ara 2023 Sal, 17:08
>>>>> tarihinde şunu yazdı:
>>>>>
>>>>>> PG 14.10
>>>>>>
>>>>>> I had initiated streaming replication on a pair of new instances from
>>>>>> Node A to Node B, and then stopped it via:
>>>>>> SELECT pg_terminate_backend(<wallreceiver pid>);
>>>>>> SELECT pg_drop_replication_slot(<slot name>);
>>>>>>
>>>>>> That seemed to work, since Node A isn't replicating anymore:
>>>>>>
>>>>>> postgres=# SELECT * from pg_replication_slots;
>>>>>> (0 rows)
>>>>>>
>>>>>> postgres=#
>>>>>> postgres=# SELECT * from pg_stat_replication;
>>>>>> (0 rows)
>>>>>>
>>>>>> Node B, unsurprisingly, doesn't have anything in pg_stat_wal_receiver.
>>>>>> $ psql --host=NodeB -xc "SELECT * FROM pg_stat_wal_receiver;"
>>>>>> (0 rows)
>>>>>>
>>>>>> And yet, Node B tries to initialize replication every 5 seconds. An
>>>>>> excerpt from the log file:
>>>>>> 2023-12-19 08:53:52.778 EST NodeB replicator 202014 initializing
>>>>>> 53300 FATAL: 53300: number of requested standby connections exceeds
>>>>>> max_wal_senders (currently 0)
>>>>>> 2023-12-19 08:53:52.778 EST NodeB replicator 202014 initializing
>>>>>> 53300 LOCATION: InitProcess, proc.c:358
>>>>>>
>>>>>>