Re: 回复:how to create index concurrently on partitioned table

Justin Pryzby <pryzby@telsasoft.com>

From: Justin Pryzby <pryzby@telsasoft.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: 李杰(慎追) <adger.lj@alibaba-inc.com>, pgsql-hackers <pgsql-hackers@lists.postgresql.org>, 曾文旌(义从) <wenjing.zwj@alibaba-inc.com>, Alvaro Herrera <alvherre@2ndquadrant.com>
Date: 2020-09-08T02:39:16Z
Lists: pgsql-hackers

Attachments

Thanks for completing and pushing the REINDEX patch and others.

Here's a rebasified + fixed version of the others.

On Tue, Sep 01, 2020 at 02:51:58PM +0900, Michael Paquier wrote:
> The REINDEX patch is progressing its way, so I have looked a bit at
> the part for CIC.
> 
> Visibly, the case of multiple partition layers is not handled
> correctly.  Here is a sequence that gets broken:
..
> This fails as follows:
> ERROR: XX000: unrecognized node type: 2139062143
> LOCATION:  copyObjectImpl, copyfuncs.c:5718

Because copyObject needed to be called within a longlived context.

Also, my previous revision failed to implement your suggestion to first build
catalog entries with INVALID indexes and to then reindex them.  Fixed.

-- 
Justin

Commits

  1. Allow CLUSTER on partitioned tables

  2. Add support for partitioned tables and indexes in REINDEX

  3. Local partitioned indexes