TODO item: WAL replay of CREATE TABLESPACE with differing directory structure

From: Patrick Krecker <pkrecker(at)gmail(dot)com>
To: PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: TODO item: WAL replay of CREATE TABLESPACE with differing directory structure
Date: 2018-02-13 21:44:34
Message-ID: CACh_hd7Syvsih+9Pq-LLpzC4Y_tfYVGsnQM1dP8MMJ+EKkLTsA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Hackers --

I am searching for a way to make a contribution to Postgres and I came
across this TODO item (I realize there has been some controversy
around the TODO list [1], and I hope that my use of it doesn't spark
another discussion about removing it altogether):

"Allow WAL replay of CREATE TABLESPACE to work when the directory
structure on the recovery computer is different from the original"

Currently it looks like tablespaces have to live inside the data
directory on the replica, notwithstanding administrator intervention
by manipulating the tablespace directory with symlinks after (or even
before?) it has been created via replay.

Is the idea behind this task to allow the master to instruct the
replica where to put the tablespace on its filesystem, so as to allow
it to live outside of the data directory without direct manipulation
of the filesystem?

If this task is a worthwhile endeavor, I would be happy to take it on.
If not, I am open to other ideas :)

Thanks,
Patrick

[1] https://www.postgresql.org/message-id/CA+TgmoZC3CyzZDY1fWChRNOY-5SBjUkB1w=C6y6JmQHtAVukjw@mail.gmail.com

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2018-02-13 22:23:21 Re: [HACKERS] Client Connection redirection support for PostgreSQL
Previous Message Peter Eisentraut 2018-02-13 21:17:13 Python 3.7 support