From: | Martijn van Oosterhout <kleptog(at)svana(dot)org> |
---|---|
To: | Mattias Kregert <mattias(at)kregert(dot)se> |
Cc: | shridhar_daithankar(at)persistent(dot)co(dot)in, PostgreSQL List <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Datatypes and performance |
Date: | 2003-07-07 11:46:13 |
Message-ID: | 20030707114613.GA30359@svana.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Mon, Jul 07, 2003 at 01:30:09PM +0200, Mattias Kregert wrote:
> In the case of disk failure, the files will probably be damaged anyway and
> then i'll have to install new hardware or format+check badblocks and then
> restore from the backup. I can't see how fsync would help in the case of
> disk crash. Without reliable raid or something i think this would be a
> catastrophic failure => get new hardware: disk/ram/computer/building
> whatever and go find the (remote) backup.
True, it doesn't help with one disk, but RAID does work.
> In the case of power failure, you'll *might* have to restore from backup
> unless you use ReiserFS or some other journalling filesystem. I use
> ReiserFS. I also have a UPS installed, just to be sure...
> Journalling FS will fix the FS problems, so the files are ok.
> PG journal will fix the PG problems so the tables will be ok.
Firstly, journalling filesystems insure the integrity of the *filesystem*,
not the files on it. So your files can still be corrupted. You could enable
full data journalling. I would imagine that would cost you more than just
enabling fsync.
Secondly, IIRC the fsync applies to the PG journals, so turning off fsync
will kill the tables in a crash.
Basically, do you care about your data?
--
Martijn van Oosterhout <kleptog(at)svana(dot)org> http://svana.org/kleptog/
> "the West won the world not by the superiority of its ideas or values or
> religion but rather by its superiority in applying organized violence.
> Westerners often forget this fact, non-Westerners never do."
> - Samuel P. Huntington
From | Date | Subject | |
---|---|---|---|
Next Message | Jean-Michel POURE | 2003-07-07 12:48:07 | Re: PostGreSql equivalents to mssql |
Previous Message | Mattias Kregert | 2003-07-07 11:30:09 | Re: Datatypes and performance |