[BUG]Update Toast data failure in logical replication

From: "tanghy(dot)fnst(at)fujitsu(dot)com" <tanghy(dot)fnst(at)fujitsu(dot)com>
To: pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: [BUG]Update Toast data failure in logical replication
Date: 2021-05-28 05:16:12
Message-ID: OS0PR01MB611342D0A92D4F4BF26C0F47FB229@OS0PR01MB6113.jpnprd01.prod.outlook.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi

I think I just found a bug in logical replication. Data couldn't be synchronized while updating toast data. Could anyone take a look at it?

Here is the steps to proceduce the BUG:
------publisher------
CREATE TABLE toasted_key (
id serial,
toasted_key text PRIMARY KEY,
toasted_col1 text,
toasted_col2 text
);
CREATE PUBLICATION pub FOR TABLE toasted_key;

------subscriber------
CREATE TABLE toasted_key (
id serial,
toasted_key text PRIMARY KEY,
toasted_col1 text,
toasted_col2 text
);
CREATE SUBSCRIPTION sub CONNECTION 'dbname=postgres' PUBLICATION pub;

------publisher------
ALTER TABLE toasted_key ALTER COLUMN toasted_key SET STORAGE EXTERNAL;
ALTER TABLE toasted_key ALTER COLUMN toasted_col1 SET STORAGE EXTERNAL;
INSERT INTO toasted_key(toasted_key, toasted_col1) VALUES(repeat('1234567890', 200), repeat('9876543210', 200));
UPDATE toasted_key SET toasted_col2 = toasted_col1;

------subscriber------
SELECT count(*) FROM toasted_key WHERE toasted_col2 = toasted_col1;

The above command is supposed to output "count = 1" but in fact it outputs "count = 0" which means UPDATE operation failed at the subscriber. Right?

I debugged and found the subscriber could receive message from publisher, and in apply_handle_update_internal function, it invoked FindReplTupleInLocalRel function but failed to find a tuple.
FYI, I also tested DELETE operation(DELETE FROM toasted_key;), which also invoked FindReplTupleInLocalRel function, and the result is ok.

Regards
Tang

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Paul Guo 2021-05-28 05:30:51 Re: Two patches to speed up pg_rewind.
Previous Message Ajin Cherian 2021-05-28 04:40:54 Re: [HACKERS] logical decoding of two-phase transactions