Re: base backup requirements for PITR - Mailing list pgsql-admin

From Albe Laurenz
Subject Re: base backup requirements for PITR
Date
Msg-id D960CB61B694CF459DCFB4B0128514C208C99F6B@exadv11.host.magwien.gv.at
Whole thread Raw
In response to Re: base backup requirements for PITR  (Gabriele Bartolini <gabriele.bartolini@2ndQuadrant.it>)
List pgsql-admin
Gabriele Bartolini wrote:
>> Gabriele, I understand that you want to spread the word,
>> but the OP's problem was a different one:

> You are right. I totally missed that. I skimmed the email (very quickly)
> and thought that in that particular context Andrew was referring to PITR
> in general.

I admit that it was a rather long e-mail.

>> He wants to create a base backup by just copying the
>> data directory of an 8.4 warm standby server.
>> There is no possibility to run pg_(start|stop)_backup,
>> but the documentation seems to suggest that it could work
>> anyway:
>> http://www.postgresql.org/docs/8.4/static/warm-standby.html#BACKUP-INCREMENTAL-UPDATED

> Yes, it is possible. I'd like to try and implement that in Barman, in
> order to automate this. I'd try and follow this path though:
> pg_start_backup() on the master, wait for the operation to be replayed
> on the standby (with pg_controldata if < 9.0 or through hot standby),
> rsync/tar from the standby, then pg_stop_backup() on the master. If you
> think of following this path yourself, please make sure you test it
> accurately (especially for production environment).

A viable alternative could be to shutdown the standby
and take an offline backup.  It can catch up with WAL
after restart.

Yours,
Laurenz Albe

pgsql-admin by date:

Previous
From: Gabriele Bartolini
Date:
Subject: Re: base backup requirements for PITR
Next
From: "Albe Laurenz"
Date:
Subject: Re: Setting up High Availability Group of PostgreSQL instances using RedHat Cluster Suite