Horton

Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Dilip Kumar <dilipbalaut@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, Heikki Linnakangas <hlinnaka@iki.fi>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-09-13T16:14:36Z
Lists: pgsql-hackers 
On Sat, Sep 11, 2021 at 12:17 AM Dilip Kumar <dilipbalaut@gmail.com> wrote:
> I agree with you, even though I think that scanning pg_class for identifying the relfilenode looks like a more sensible thing to do than directly scanning the file system, we need to consider one point that, now also in current system  (in create database) we are scanning the directory for copying the file so instead of copying them directly we need to logically identify the relfilenode and then copy it block by block, so maybe this approach will not make anyone unhappy because it is not any worse than the current system.

So, I agree. If we can't get agreement on this approach, then we can
do that, and as you say, it's no worse than what we are doing now. But
I am just trying to lay out my view of why I think that's not as good
as this.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

  1. When using the WAL-logged CREATE DATABASE strategy, bulk extend.

  2. Avoid using a fake relcache entry to own an SmgrRelation.

  3. Fix data-corruption hazard in WAL-logged CREATE DATABASE.

  4. initdb: When running CREATE DATABASE, use STRATEGY = WAL_COPY.

  5. Simplify a needlessly-complicated regular expression.

  6. In 020_createdb.pl, change order of command-line arguments.

  7. Add new block-by-block strategy for CREATE DATABASE.

  8. Fix replay of create database records on standby

  9. Refactor code for reading and writing relation map files.

  10. Replace RelationOpenSmgr() with RelationGetSmgr().

  11. Refactor the fsync queue for wider use.