Re: Locking in Inheritance partition - Mailing list pgsql-admin

From Gaurav Anand
Subject Re: Locking in Inheritance partition
Date
Msg-id CAMZ99ndZexsNuEu1CWZeB6GVPdoivHRr0pY2UmS7bUE1ws9YrA@mail.gmail.com
Whole thread Raw
In response to Re: Locking in Inheritance partition  (Thomas Kellerer <shammat@gmx.net>)
List pgsql-admin
Thank you Thomas.

Upgrade is not planned in near future.

With pg10, we added advisory lock. but when it comes to detach attach we are still struggling.



On Fri, 10 Dec 2021 at 4:07 PM, Thomas Kellerer <shammat@gmx.net> wrote:
Gaurav Anand schrieb am 10.12.2021 um 07:35:
> We have a declarative partition on tables in our database which is
> PG10. We face issues as the table becomes locked during attach and
> detach and cannot be used during the event.
>
> We have a master and slave setup and application query points to
> Slave but whenever we perform attach and detach of partitions on
> Master the locks propagates to slave and causes Application to spin.
>
> Kindly suggest if  Inheritance partition will help here or any other
> way I can avoid this lock.

I don't think inheritance would change the locks this requires.

A better solution might be to upgrade to 14 where both attach
and detach are non-blocking.

Postgres 12 introduced non-blocking attach and Postgres 14
introduced non-blocking detach

Thomas


--

Thanks.

Regards,
Gaurav Anand

logo

This communication is confidential and subject to and governed by Saama’s Electronic Communications Disclaimer.

 

pgsql-admin by date:

Previous
From: "Michel SALAIS"
Date:
Subject: RE: Very slow 101-feeling design/query..
Next
From: Wells Oliver
Date:
Subject: Re: Very slow 101-feeling design/query..