Re: Turning slave into a master - PostgreSQL 9.2 - Mailing list pgsql-general

From Patrick B
Subject Re: Turning slave into a master - PostgreSQL 9.2
Date
Msg-id CAJNY3iu2GNVPQ14RNzL2SNyEf0iQO9xZKcBwYQ+-gsvjJV5eyw@mail.gmail.com
Whole thread Raw
In response to Re: Turning slave into a master - PostgreSQL 9.2  ("David G. Johnston" <david.g.johnston@gmail.com>)
Responses Re: Turning slave into a master - PostgreSQL 9.2
List pgsql-general


2016-11-01 10:33 GMT+13:00 David G. Johnston <david.g.johnston@gmail.com>:
On Mon, Oct 31, 2016 at 1:46 PM, Patrick B <patrickbakerbr@gmail.com> wrote:
Hi guys,

I got a test server, let's call it test01.

The test01 has a basebackup from the master.
I want to turn test01 into a master. It doesn't need to catch up with the wal_files, because I don't need it to be up-to-date.

So what I did is:

- Replaced /var/lib/pgsql/9.2/data/ with the basebackup
- Created recovery.conf:
restore_command = 'cp /var/lib/pgsql/wal_archive/%f %p'
recovery_target_timeline = 'latest'
standby_mode = off
trigger_file = '/tmp/pg_failover_trigger' 
- touch /tmp/pg_failover_trigger
- service postgresql start

And then postgres starts recovering the wal_files. But I don't want that.. as I don't need a up-to-date

Is the wal_files required anyway?

 
​"...has a basebackup from the master" - the answer to your question depends greatly on the detail behind that sentence.

IIRC, unless you know that the data directory is consistent - because the database was offline at the time of the backup - at least some WAL will probably be required to bring the inconsistent backup data directory to a known good state (i.e., post-checkpoint).

David J.
 


I see...


as I'm recovering a slave and then turning it into a master, that's why the wal_files are required.

Thanks!
Patrick

pgsql-general by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: Turning slave into a master - PostgreSQL 9.2
Next
From: Melvin Davidson
Date:
Subject: Re: Way to quickly detect if database tables/columns/etc. were modified?