Re: Slony and triggers on slavenodes. - Mailing list pgsql-general

From Achilleas Mantzios
Subject Re: Slony and triggers on slavenodes.
Date
Msg-id ddcc711f-0c46-089b-6f27-31cb92e59e2a@matrix.gatewaynet.com
Whole thread Raw
In response to Slony and triggers on slavenodes.  (Gustavsson Mikael <mikael.gustavsson@smhi.se>)
Responses SV: Slony and triggers on slavenodes.  (Gustavsson Mikael <mikael.gustavsson@smhi.se>)
List pgsql-general
On 18/2/19 1:09 μ.μ., Gustavsson Mikael wrote:
P {margin-top:0;margin-bottom:0;}
Hi,

I have a Postgresql 11 and Slony 2.2.7 setup with one master and multiple slave nodes. On one of the slaves I want a user defined trigger to fire on insert. The trigger fires a notify function.

Now to the problem.

If I create the trigger it looks like it is enabled on the slave but it do not fire.

CREATE TRIGGER <trigger_name>
    AFTER INSERT
    ON <table_name>
    FOR EACH ROW
    EXECUTE PROCEDURE <function_name>;

If i try to enable it explicit with

ALTER TABLE <table_name>
ENABLE ALWAYS TRIGGER <trigger_name>;

or

ALTER TABLE <table_name>
ENABLE REPLICA TRIGGER <trigger_name>;

and then look at the table definition the trigger is disabled.

How? can you show the output of \d ? or pg_dump -x --schema-only -t <table_name> ?


Is this expected behavior in slony?
I understand that this may be tricky in replication context and that the default behavior is that triggers on slaves are disabled.

We have plans to move to logical replication which should solve this I assume, but it does not fit the schedule at the moment.

KR
Mikael Gustavsson



-- 
Achilleas Mantzios
IT DEV Lead
IT DEPT
Dynacom Tankers Mgmt

pgsql-general by date:

Previous
From: Gustavsson Mikael
Date:
Subject: Slony and triggers on slavenodes.
Next
From: Jahwan Kim
Date:
Subject: Cannot vacuum even in single-user mode after xidStopLimit is reached