Re: locking [user] catalog tables vs 2pc vs logical rep

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: pgsql-hackers(at)postgresql(dot)org, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, Markus Wanner <markus(dot)wanner(at)enterprisedb(dot)com>
Subject: Re: locking [user] catalog tables vs 2pc vs logical rep
Date: 2021-05-24 05:03:12
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

On Mon, Feb 22, 2021 at 02:28:47PM -0800, Andres Freund wrote:
> Perhaps all that we need to do is to disallow 2PC prepare if [user]
> catalog tables have been locked exclusively? Similar to how we're
> disallowing preparing tables with temp table access.

At least for anything involving critical relations that get loaded at
startup? It seems to me that if we can avoid users to get them
completely locked out even if they run the operation on an object
they own, that would be better than requiring tweaks involving
pg_resetwal or equal to rip of the 2PC transaction from existence.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Dilip Kumar 2021-05-24 05:04:36 Re: Race condition in recovery?
Previous Message Michael Paquier 2021-05-24 04:50:13 Re: Re: Parallel scan with SubTransGetTopmostTransaction assert coredump