Re: [GENERAL] Large database

From: Bruce Momjian <maillist(at)candle(dot)pha(dot)pa(dot)us>
To: Jim Mercer <jim(at)reptiles(dot)org>
Cc: Ole Gjerde <gjerde(at)icebox(dot)org>, gd(at)pinmail(dot)com, pgsql-general(at)postgreSQL(dot)org
Subject: Re: [GENERAL] Large database
Date: 1999-08-17 17:03:54
Message-ID: 199908171703.NAA01869@candle.pha.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

> > > 2. Is it possible for database tables to span files on more than one disk
> > > partition? I'm aware that it splits files into 2GB chunks
> > > automatically, but is it possible for these files to reside on more
> > > than one partition?
> >
> > Easiest way to do this would be to put all your drives in a RAID.. Either
> > just striping or RAID 5 if you need the redundancy(probably).
>
> oddly enough, we had 2GB chunks with 6.4, but now with 6.5.1 (on FreeBSD)
> we are seeing 1GB chunks:
>
> -rw------- 1 pgsql wheel 1073741824 Aug 15 08:23 gt
> -rw------- 1 pgsql wheel 1073741824 Aug 15 08:29 gt.1
> -rw------- 1 pgsql wheel 626548736 Aug 17 12:29 gt.2
> -rw------- 1 pgsql wheel 302964736 Aug 17 12:24 gt_callid_key
> (we've had as much as gt.7)
>
> doesn't seem to have broken anything.

Yes, this was done because we found some OS's couldn't handle chunks of
exactly 2G, so we reduced it, and 1G seemed a good number.

>
> it would be nice if vacuum could figure out if an extent is no longer
> in use, and delete it.
>
> currently, we need to dump the table, then nuke it and restore it.

We tried removing the extra extent, but the other backends needed it
around because there was no way to tell them that the extra extent file
descriptor was useless. You could just stop the postmaster, and delete
the zero-length extents, if you really wanted to. Do the extra extent
files cause problems for you? If so, we can start to look at some
postmaster cleanup of those.

>
> > > 3. What's the biggest known database currently running on Postgresql?
> >
> > No clue.. The biggest one I'm running right now is about 5GB which isn't
> > all that much.. I'm sure there are much bigger ones out there..
>
> my table varies in size from 1 - 7 gig.
> it has millions of records with a single index.
>
> works for the most part, the problems we have are usually related to the
> hardware crashing, and since we run postgres without the sync flag, well,
> things can get corrupted.

--
Bruce Momjian | http://www.op.net/~candle
maillist(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Shawn Pursley 1999-08-17 17:16:50 Populating a Class with Instances
Previous Message Blks 1999-08-17 17:03:52 Never ending (for) loop