Re: PITR Setup Using Hot StandBy - Mailing list pgsql-admin
I'm still not sure I fully understand what it is you're trying to accomplish.
Using hot-standby streaming replication, you can have a slave up to date with the master and available for promotion as well as select-only queries.
If you're trying to setup multiple slaves, this is also possible.
I have a production server (call it db-prod). I have a training/demo server (also at my data center -- call it db-back). I have a 3rd server at my HQ (and a 4th and 5th in my DR data center).
DB-prod runs my paying customers
DB-back runs my test/training sites (different port) and replicates DB-prod
DB-HQ runs my internal stuff (standard port) and replicates DB-back. It replicates BOTH DB-back AND DB-prod FROM DB-back. It does not touch the actual production server
DB-DR replicates DB-prod
DB-DR-back replicates DB-back (both clusters)
I accomplish this using streaming replication in PG 9.x.
Sent: Tuesday, May 24, 2016 12:21 PM
To: Scott Whitney; Dave Johansen
Cc: pgsql-admin@postgresql.org
Subject: RE: [ADMIN] PITR Setup Using Hot StandBy
Hi,
Basically I do not want to touch my Master while performing PITR.
So here what I am trying to achieve now in my development environment.
1. Take base backup from slave
2. Do WAL archiving from master
3. Try to up postgresql using base backup from slave and wal archives from master.
I don’t know, if this will work or not.
But my final goal is to setup entire PITR process using base backup from slave.
Thanks & Regards,
Vicky Soni
Database Administrator
From: Scott Whitney [mailto:scott@journyx.com]
Sent: 24 May 2016 21:18
To: Dave Johansen <davejohansen@gmail.com>; Vicky Soni - Quipment India <vicky.soni@quipment.nl>
Cc: pgsql-admin@postgresql.org
Subject: Re: [ADMIN] PITR Setup Using Hot StandBy
Hi, Dave.
It depends on what you're trying to do.
If you want a slave up in real-time replication mode, that's by far the simplest setup.
The concept is:
1) start a backup
select pg_start_backup("myslave"),true);
2) Rsync (or tar) your pg data directory over to your slave.
3) select pg_stop_backup();
4) Tell your slave how to connect to the master
Start your slave.
This is a pretty straight-forward link for this:
https://opensourcedbms.com/dbms/how-to-do-point-in-time-recovery-with-postgresql-9-2-pitr-3/
| opensourcedbms.com This post describes how to do point in time recovery with PostgreSQL 9.2 : PITR. Point in time recovery (PITR) simply means restoring data upto certain point in time. |
If you want to do something "clever," like offset how far behind the slave is, that's more complex.
From: pgsql-admin-owner@postgresql.org <pgsql-admin-owner@postgresql.org> on behalf of Dave Johansen <davejohansen@gmail.com>
Sent: Tuesday, May 24, 2016 10:31 AM
To: Vicky Soni - Quipment India
Cc: pgsql-admin@postgresql.org
Subject: Re: [ADMIN] PITR Setup Using Hot StandBy
On Tue, May 24, 2016 at 6:05 AM, Vicky Soni - Quipment India <vicky.soni@quipment.nl> wrote:
Hello Friends,
I need to setup PITR. Using master, I am able to do wal archiving.
But using slave I am not able to archive WAL files.
Can anyone please guide me, how to setup PITR using Hot StandBy?
I have setup LogShpping + Streaming Replication in my dev environment, now using this slave I want to setup PITR.
I have already setup following parameters:
wal_level = hot_standby
archive_mode = on
archive_command = 'copy "c:\\Program Files\\PostgreSQL\\9.4\\data\\%p" "\\\\192.168.170.221\\Primary_WAL_Archive\\%f"'
max_wal_senders = 3
wal_keep_segments = 8
hot_standby = on
Any help is appreciated.
http://www.postgresql.org/docs/9.2/static/continuous-archiving.html
The standby/slave is read only and does not generate WAL files (it only consumes them to stay in sync with the master).
Journyx, Inc.
7600 Burnet Road #300
Austin, TX 78757
www.journyx.com
p 512.834.8888
f 512-834-8858
Do you receive our promotional emails? You can subscribe or unsubscribe to those emails at http://go.journyx.com/emailPreference/e/4932/714/
Attachment
pgsql-admin by date: