Re: Delay locking partitions during INSERT and UPDATE

From: David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>
To: John Naylor <jcnaylor(at)gmail(dot)com>
Cc: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Delay locking partitions during INSERT and UPDATE
Date: 2019-01-20 21:04:22
Message-ID: CAKJS1f9beiqzppvh1e7tNeY-r186pX2V57N8H8MU_thDzSVH9Q@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sat, 19 Jan 2019 at 12:05, John Naylor <jcnaylor(at)gmail(dot)com> wrote:
>
> On 11/22/18, David Rowley <david(dot)rowley(at)2ndquadrant(dot)com> wrote:
> > If required, such operations could LOCK TABLE the top partitioned
> > table to block the DML operation. There's already a risk of similar
> > deadlocks from such operations done on multiple separate tables when
> > the order they're done is not the same as the order the tables are
> > written in a query, although, in that case, the window for the
> > deadlock is likely to be much smaller.
>
> Is this something that would need documentation anywhere?

Probably at least the release notes. I'm unsure where else to mention
it. I don't feel the workaround of using LOCK TABLE is special to
this case. The patch does, however, make it a possible requirement for
performing DDL on individual partitions where it was not previously.

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Fabien COELHO 2019-01-20 21:54:41 Re: PSA: we lack TAP test coverage on NetBSD and OpenBSD
Previous Message Kim Rose Carlsen 2019-01-20 20:51:22 Postgres doesn't remove useless join when using partial unique index