Re: OK to put temp tablespace on volatile storage or to omit it from backups?

From: Christophe Pettus <xof(at)thebuild(dot)com>
To: pgsql-general list <pgsql-general(at)postgresql(dot)org>
Subject: Re: OK to put temp tablespace on volatile storage or to omit it from backups?
Date: 2013-05-01 16:17:43
Message-ID: 5A52C82F-78AD-4A62-8CC5-B93E38D2815C@thebuild.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general


On Apr 30, 2013, at 11:34 PM, Yang Zhang wrote:

> In other words, I guess, I'm asking because of Xof's comment on that blog post:
>
> "That’s true if you recreate the PG_ directory in the tablespace;
> otherwise, you get the error:
> reindexdb: reindexing of database "test" failed: ERROR: could not
> create directory "pg_tblspc/69944/PG_9.2_201204301/61884": No such
> file or directory
> However, that’s not a guaranteed feature of PostgreSQL, and I would
> never rely on it."

Well, here's the core situation: If the underlying storage for a tablespace disappears, you have to patch up the database by recreating the directory in order for the database to be operational again. This isn't a documented PostgreSQL API; it's a disaster recovery procedure. It works now, it probably will work in the future, but I don't think there's any guarantee that the procedure that works today will work tomorrow.

Thus, I really don't recommend making an operational decision that the lost of a tablespace's storage is considered something routine.

That being said, you can make it work today if it is critical that it do so.

--
-- Christophe Pettus
xof(at)thebuild(dot)com

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Andrew Hastie 2013-05-01 16:34:43 Re: zLinux Load Testing Experience
Previous Message Christophe Pettus 2013-05-01 16:13:27 Re: OK to put temp tablespace on volatile storage or to omit it from backups?