Horton

Re: ATTACH/DETACH PARTITION CONCURRENTLY

Peter Eisentraut <peter.eisentraut@2ndquadrant.com>

View thread View on postgr.es
From: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
To: Andres Freund <andres@anarazel.de>, David Rowley <david.rowley@2ndquadrant.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Simon Riggs <simon@2ndquadrant.com>
Date: 2018-08-09T18:57:35Z
Lists: pgsql-hackers 
On 07/08/2018 15:29, Andres Freund wrote:
> I don't think that solves the problem that an arriving relcache
> invalidation would trigger a rebuild of rd_partdesc, while it actually
> is referenced by running code.

The problem is more generally that a relcache invalidation changes all
pointers that might be in use.  So it's currently not safe to trigger a
relcache invalidation (on tables) without some kind of exclusive lock.
One possible solution to this is outlined here:
<https://www.postgresql.org/message-id/CA+TgmobtmFT5g-0dA=vEFFtogjRAuDHcYPw+qEdou5dZPnF=pg@mail.gmail.com>

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Commits

  1. Allow ATTACH PARTITION with only ShareUpdateExclusiveLock.

  2. Change lock acquisition order in expand_inherited_rtentry.

  3. Move code for managing PartitionDescs into a new file, partdesc.c

  4. Remove more redundant relation locking during executor startup.

  5. Add assertions that we hold some relevant lock during relation open.

  6. Try to acquire relation locks in RangeVarGetRelid.