Обсуждение: Some silly questions
I was just wondering about a few options that I would think might be pretty useful. They might exist, but then, I don't know how to get it done... Here's my list: - multi-volume database. Is it possible to split the complete database so it will use several disks? One reason might be load-balancing of the disks, another might be a very full disk... there may be even more... - If multi-voluming exists, is it possible to make certain specific tables can get in a special 'area', so you'd know on which disk it would be? - Every database system has a ROLLBACK function, also known as 'before-imaging'. Postgres no doubt has this technique as well? How (which file(s)) is this done in Postges, and how can you manage the size of these before images? - In a situation of a crash, a rollback is a MUST. A rollforward is desirable... does postgres offer this functionality as well? (I guess it does) and the same question as the before-image arises as with this so-called after-imaging. that's it for now... I'm just pretty curious I think... Oh... and does anyone know something about the relation Postgres and Progress? the names seem so similar... and... do my questions sound stupid? ;-) Michiel Lange
Hello list: Has anybody know how to see all user function, and how to see the code? Not with pg_access. I mean command line in psql? Thanks in advance.
SELECT * from pg_proc where proname='<name of procedure>'; > Hello list: > > Has anybody know how to see all user function, and how to see the code? Not with pg_access. I > mean command line in psql? > > Thanks in advance. > > ---------------------------(end of broadcast)--------------------------- TIP 4: Don't 'kill -9' > the postmaster ----------------------------------------- Get your free web based email at trade-india.com. "India's Leading B2B eMarketplace.!" http://www.trade-india.com/
"Michiel Lange" <michiel@minas.demon.nl> wrote in message news:apn9vs$gou$1@news.hub.org... > I was just wondering about a few options that I would think might be pretty > useful. They might exist, but then, I don't know how to get it done... > > Here's my list: > - multi-volume database. Is it possible to split the complete database so it > will use several disks? One reason might be load-balancing of the disks, > another might be a very full disk... there may be even more... > - If multi-voluming exists, is it possible to make certain specific tables > can get in a special 'area', so you'd know on which disk it would be? Well, if you take a look in the directory /var/lib/pgsql/data/base (this path depend on your distribution ) you can see a lot of directory (numbers) and at this level each number is a Database inside that directory you have others file and each is a different object of the DB ( index, table etc) you can create link on different disk if you want, take care that Postgres don't follow the links so if you delete and create a table or an index that object is recreated on the original location. Ciao Gaetano
On Wed, Oct 30, 2002 at 01:36:28AM +0100, Michiel Lange wrote:
> I was just wondering about a few options that I would think might be pretty
> useful. They might exist, but then, I don't know how to get it done...
>
> Here's my list:
> - multi-volume database. Is it possible to split the complete database so it
> will use several disks? One reason might be load-balancing of the disks,
> another might be a very full disk... there may be even more...
> - If multi-voluming exists, is it possible to make certain specific tables
> can get in a special 'area', so you'd know on which disk it would be?
There have been several suggestions to implement this functionality,
but it's been a relatively low priority, it seems. You can sort of
mimic this with symlinks, but it's a pain to do. My suggestion is to
throw money at a RAID configuration, instead. We've had very good
luck with RAID 1+0.
> - Every database system has a ROLLBACK function, also known as
> 'before-imaging'. Postgres no doubt has this technique as well? How (which
> file(s)) is this done in Postges, and how can you manage the size of these
> before images?
ROLLBACK is a feature of the transaction model in Postgres. You
already have it; you need to initiate a transaction to get to it.
Postgres does this with the WAL. You _do_ want to put the WAL files
(which are in pg_xlog in your data directory) on another disk. You
may do this with a symlink; SHUT DOWN YOUR DATABASE before you do it!
Check out the WAL section of the manual for more details on how to
use and configure the facility.
> - In a situation of a crash, a rollback is a MUST. A rollforward is
> desirable... does postgres offer this functionality as well? (I guess it
> does) and the same question as the before-image arises as with this
> so-called after-imaging.
There is automatic recovery in the event of a crash, but you can't
get point-in-time recovery yet. It's in the hopper for 7.4,
apparently.
> Oh... and does anyone know something about the relation Postgres and
> Progress? the names seem so similar...
Not as far as I know. There is an historical relationship between
PostgreSQL, Postgres (which is actually different from PostgreSQL,
and had a commercial variant in Illustra), and Ingres. See the
history section of the manual for details.
> and... do my questions sound stupid? ;-)
No, but you might want to have a look at the manuals: there's a lot
of info there.
A
--
----
Andrew Sullivan 204-4141 Yonge Street
Liberty RMS Toronto, Ontario Canada
<andrew@libertyrms.info> M2P 2A8
+1 416 646 3304 x110