Re: BUG #8450: pg_basebackup blocks until WAL archiving successful

From: Stuart Bishop <stuart(at)stuartbishop(dot)net>
To: Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #8450: pg_basebackup blocks until WAL archiving successful
Date: 2013-09-25 19:49:48
Message-ID: CADmi=6P5H5OFAqzFyi_Y=y+xRQ353K=b6i+7N5TNhfAezNeaWg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

On Mon, Sep 23, 2013 at 3:33 PM, Heikki Linnakangas

> I can see why you'd want that, but it seems equally problematic to let
> pg_basebackup return, when the WAL files haven't been archived yet and you
> therefore don't in fact have valid, restorable backup yet. Have you
> considered using the --xlog-method=stream option, to include the WAL files
> in the backup? That will make your backups somewhat larger, as the WAL files
> are included, but in that mode pg_basebackup won't wait for the archival and
> the backup will be restorable even if archive_command is failing.

I'm supporting PG 9.1 at the moment so cannot rely on
--xlog-method=stream. I agree that the current behavior is for most
use cases better, and I think that the behavior I want should be
explicitly enabled with an option.

--
Stuart Bishop <stuart(at)stuartbishop(dot)net>
http://www.stuartbishop.net/

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Christoph Berg 2013-09-26 12:59:30 pg_upgrade 9.0->9.2 failure: Mismatch of relation OID in database
Previous Message Alvaro Herrera 2013-09-25 14:26:19 Re: BUG #8470: 9.3 locking/subtransaction performance regression