[pgpool-general: 9448] Re: Clarification on query results cache visibility

[Achilleas Mantzios - cloud]

On 4/29/25 11:03, Tatsuo Ishii wrote:
>> Thank you a lot for such a quick response ! I applied, and re-run the
>> test, it works correctly.
>>
>> We will keep testing as in live conditions, with wildfly , etc.
>>
>> I don't understand two things :
>>
>> - I also tested a non-prepare statement version of the java test, when
>> - run against 5432 (pgsql) the pgsql logs showed only Execute(s) as
>> - expected. When run via 9999 (pgpool),  the pgpool logs showed :
>> - Parse/Bind/Execute flow as if they were Prepared Statements.
> PostgreSQL's log_statement does show Execute messages, but does not
> show Parse/Bind messages. On the other hand Pgpool's log_statement
> shows Parse/Bind/Execute (plus other extended query protocol messages
> including Describe and Sync).
You are correct! I have to set log_min_duration_statement to 0 in order 
to get the Parse/Bind as well!
>> - Why didn't the problem manifest itself during the sql-only tests, why
>> - we could not replicate it with plain SQL?
> Because the bug is only related to extended query protocol.
> "partial_fetch" flag is only set to true while processing execute
> message, but execute message does not exist in simple query
> protocol. The bug is only exhibited when the partial_fetch flag is
> true.
Thank you! it makes perfect sense !
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS K.K.
> English: http://www.sraoss.co.jp/index_en/
> Japanese:http://www.sraoss.co.jp