Re: PITR question with base backup

Lists: pgsql-admin
From: "Scott Whitney" <swhitney(at)journyx(dot)com>
To: <pgsql-admin(at)postgresql(dot)org>
Subject: PITR question with base backup
Date: 2008-10-21 20:02:39
Message-ID: 20081021200239.592B47E42AE@mail.int.journyx.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-admin

I'm in the process of testing PITR recovery, and I have an issue.

My "data" directory is 30GB. Not huge, but it certainly takes awhile to tar
up.

My understanding is:
a) pg_start_backup
b) tar up
c) pg_stop_backup
d) restore tar file

The problem is that I create databases pretty regularly. Let's say I create
3 in a week. I'm not looking forward to going to my colo, grabbing the 20ish
GB tgz file and restoring it 3 times per week. I'd rather do that dance
monthly or quarterly and rely on the WALs in the interim.

So...can I:

a) pg_start_backup
b) tar up my data dir
c) pg_stop_backup
d) restore
e) WALs are all happy
f) Day 2 - create new db
g) pg_start_backup
h) JUST grab the data/base directory for the new db
i) pg_stop_backup

move on?

What about "global" or "pg_clog" or "pg_multixact" or "pg_tblspc" and
"pg_twophase?"

I don't _think_ I would ever want clog, but maybe I'm wrong.

Is this possible, or do I absolutely need an entire new backup of the whole
data dir?


From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: Scott Whitney <swhitney(at)journyx(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: PITR question with base backup
Date: 2008-10-21 20:09:51
Message-ID: 20081021200951.GC4001@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-admin

Scott Whitney wrote:

> The problem is that I create databases pretty regularly. Let's say I create
> 3 in a week. I'm not looking forward to going to my colo, grabbing the 20ish
> GB tgz file and restoring it 3 times per week. I'd rather do that dance
> monthly or quarterly and rely on the WALs in the interim.

What's the actual problem? Database creation is also replayed via WAL.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


From: "Scott Whitney" <swhitney(at)journyx(dot)com>
To: <pgsql-admin(at)postgresql(dot)org>
Subject: Re: PITR question with base backup
Date: 2008-10-21 20:29:52
Message-ID: 20081021202952.C95047E4246@mail.int.journyx.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-admin

It is, is it? I was completely under the impression that it was not. Don't
ask me where I got that impression. :)

No problem whatsoever, in that case!

Thanks for clearing up my inability to comprehend documentation...

-----Original Message-----
From: Alvaro Herrera [mailto:alvherre(at)commandprompt(dot)com]
Sent: Tuesday, October 21, 2008 3:10 PM
To: Scott Whitney
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: [ADMIN] PITR question with base backup

Scott Whitney wrote:

> The problem is that I create databases pretty regularly. Let's say I
create
> 3 in a week. I'm not looking forward to going to my colo, grabbing the
20ish
> GB tgz file and restoring it 3 times per week. I'd rather do that dance
> monthly or quarterly and rely on the WALs in the interim.

What's the actual problem? Database creation is also replayed via WAL.

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support


From: "Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com>
To: "Scott Whitney" <swhitney(at)journyx(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: PITR question with base backup
Date: 2008-10-22 02:13:44
Message-ID: dcc563d10810211913r75736d25nebb8f3258930762c@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-admin

On Tue, Oct 21, 2008 at 2:29 PM, Scott Whitney <swhitney(at)journyx(dot)com> wrote:
> It is, is it? I was completely under the impression that it was not. Don't
> ask me where I got that impression. :)
>
> No problem whatsoever, in that case!
>
> Thanks for clearing up my inability to comprehend documentation...

This is what I like so much about PostgreSQL. Most of the surprises
are of the "oh wow! That's cool" Not the "oh shit!" kind. :)