Message-ID: <dd5c48d4-8962-d1c9-e71a-6bda23b86401@enterprisedb.com>
Date: Fri, 17 Nov 2023 15:40:35 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.15.1
Subject: Re: Logical replication is missing block of rows when sending initial
 sync?
Content-Language: en-US
To: "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>,
 "'depesz@depesz.com'" <depesz@depesz.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>,
 Kyotaro Horiguchi <horikyota.ntt@gmail.com>,
 "pgsql-bugs@lists.postgresql.org" <pgsql-bugs@lists.postgresql.org>
References: 
 <TYAPR01MB5866C2BE11AD4FBDC0AB46F0F5A1A@TYAPR01MB5866.jpnprd01.prod.outlook.com>
 <20231102.101713.170587046871054404.horikyota.ntt@gmail.com>
 <ZUOGohG9DArmXNvV@depesz.com>
 <CAA4eK1KDMp_hCFZo=KsbRUOn+JX70Te5BJcTKwfRrt2x++aXXA@mail.gmail.com>
 <ZUTh6bnNFHo35/Go@depesz.com>
 <a266b461-e340-10c6-d511-40f9edb37d28@enterprisedb.com>
 <ZUTmEi8mEoDl93Vm@depesz.com> <ZUt9FOueePOl2ZUp@depesz.com>
 <ZUuU2uKd4fW+8Ctp@depesz.com>
 <2a638962-7a62-abd6-d9a4-3c322ec2a06d@enterprisedb.com>
 <ZU4XkdOPCthiyN7L@depesz.com>
 <TYAPR01MB5866C96DDABDB0D7F2850972F5B1A@TYAPR01MB5866.jpnprd01.prod.outlook.com>
From: Tomas Vondra <tomas.vondra@enterprisedb.com>
In-Reply-To: 
 <TYAPR01MB5866C96DDABDB0D7F2850972F5B1A@TYAPR01MB5866.jpnprd01.prod.outlook.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Archived-At: 
 <https://www.postgresql.org/message-id/dd5c48d4-8962-d1c9-e71a-6bda23b86401%40enterprisedb.com>
Precedence: bulk

On 11/15/23 02:41, Hayato Kuroda (Fujitsu) wrote:
> Dear Tomas, Depesz,
> 
>>
>>> I'll try reproducing this locally over the weekend. Should I use the
>>> test_1030.sh script that you shared a week ago, or do I need to do
>>> something more?
>>
>> That question is probably not to me, but to Hayato Kuroda
> 
> Just FYI - As I posted, this script could not reproduce the issue. We may use the same table definition
> but I have not tried yet. I will let you know if I succeeded to reproduce.
> 

FYI I think I've reproduced the issue (with a different script), I've
started a thread on pgsql-hackers:

https://www.postgresql.org/message-id/de52b282-1166-1180-45a2-8d8917ca74c6%40enterprisedb.com

Obviously, I can't be sure it's exactly the same issue, but the symptoms
seem to match what Depesz reported (gaps in data etc.). I don't know
what the root cause is, or fix. But there's more info (and scripts) in
the hackers thread.

regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company