Re: [HACKERS] Problems with >2GB tables on Linux 2.0

A friend of mine passed on an old DBA trick which
might be useful here (albeit quite the hack).  For
optimization purposes, a DBA might not only want to
put certain datafiles which composed tablespaces
(in the Oracle sense) on a striped partition, but
he might also want either to place a single table on
a RAID striped partition, or, if RAID was unavailable,
split the table across multiple disks on multiple
controllers.  This was done for a table, say z as
follows:

CREATE TABLE x...
CREATE TABLE y...

[ perform some inserts into x and y from z]

DROP TABLE z

CREATE VIEW z AS SELECT * FROM x UNION SELECT * FROM y;

Using this little trick will get you around the 2Gig
limit for now, although inserting and updating
through the view might cause problems, or might
require some use of RULES to ensure proper inserting,
updating, deleting etc.  It typically was done on
largely huge, static tables.


Hope this helps (as a hack-around) for now,

Marcus Mascari (mascarim@yahoo.com)

---The Hermit Hacker <scrappy@hub.org> wrote:
>
> On Sat, 6 Feb 1999, Thomas Reinke wrote:
>
> > I may be dating myself really badly here, but isn't there a hard
limit
> > on
> > the file system at 2Gig? I thought the file size attribute in Unix
is
> > represented as a 32 bit signed long, which happens to be a max value
> > of 2147483648. If I'm right, it means the problem is fundamentally
> > with the file system, not with PostGres, and you won't solve this
> > unless the os supports larger files.
>
> PostgreSQL has internal code that is supposed to automagically break
up a
> table into 2gb chunks so that thsi isn't a problem...
> >
> > gjerde@icebox.org wrote:
> > >
> > > Hi,
> > > I saw a message a couple of weeks ago from someone having
problems with
> > > larger than 2GB tables.  I have similar problems.
> > >
> > > PostgreSQL: anon-cvs as of today (2/5/1999)
> > > OS: Redhat Linux 5.2 (running 2.0.35)
> > >
> > > I created a database called mcrl, and a table called mcrl3_1.
> > > I copied in a set of 450MB of data twice(which comes to pg file
size of
> > > 2.4GB or so).
> > >
> > > When it hit 2GB I got this message:
> > >   mcrl=> copy mcrl3_1 FROM '/home/gjerde/mcrl/MCR3_1.txt';
> > >   ERROR:  mcrl3_1: cannot extend
> > >
> > > The table file looks like this:
> > > [postgres@snowman mcrl]$ ls -l mcrl*
> > > -rw-------   1 postgres postgres 2147482624 Feb  5 16:49 mcrl3_1
> > >
> > > It did NOT create the .1 file however, which I did see when I
tried this
> > > on 6.4.2(but still didn't work).
> > >
> > > I looked around in the code(specifically
src/backend/storage/smgr/*.c),
> > > but couldn't figure too much of it out.  I'll have to figure out
how
> > > postgres handles the database files first..
> > >
> > > Hope this helps,
> > > Ole Gjerde
> >
> > --
> > ------------------------------------------------------------
> > Thomas Reinke                            Tel: (416) 460-7021
> > Director of Technology                   Fax: (416) 598-2319
> > E-Soft Inc.                         http://www.e-softinc.com
> >
>
> Marc G. Fournier
> Systems Administrator @ hub.org
> primary: scrappy@hub.org           secondary:
scrappy@{freebsd|postgresql}.org
>
>
>

_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com