The order of queues in row lock is changed (not FIFO) - Mailing list pgsql-hackers

From Ryo Yamaji (Fujitsu)
Subject The order of queues in row lock is changed (not FIFO)
Date
Msg-id TYAPR01MB607370A5022B8854EFDC1BA58AAD9@TYAPR01MB6073.jpnprd01.prod.outlook.com
Whole thread Raw
Responses Re: The order of queues in row lock is changed (not FIFO)  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Hi all,

In pgql-general, I reported that the queue order changed in
the following cases. [1]
・Multiple sessions request row locks for the same tuple
・Update occurs for target tuple

I would like to hear the opinion of experts on whether it is a
specification or a bug.
I think row locking is a FIFO specification, using tuple header
and lock managers. Therefore, I think that the above is a bug.

[1]
https://www.postgresql.org/message-id/TYAPR01MB6073506ECCD7B8F51DA807F68A0F9@TYAPR01MB6073.jpnprd01.prod.outlook.com


Regerds.



pgsql-hackers by date:

Previous
From: Harinath Kanchu
Date:
Subject: Re: LOG: invalid record length at : wanted 24, got 0
Next
From: Michael Paquier
Date:
Subject: Re: Simplify standby state machine a bit in WaitForWALToBecomeAvailable()