Re: Multiple disks

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "scott(dot)marlowe" <scott(dot)marlowe(at)ihs(dot)com>
Cc: hal <hal(at)cc(dot)usu(dot)edu>, pgsql-admin(at)postgresql(dot)org
Subject: Re: Multiple disks
Date: 2004-04-09 16:33:02
Message-ID: 26548.1081528382@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

"scott.marlowe" <scott(dot)marlowe(at)ihs(dot)com> writes:
> The next way is to individually link thinks like indexes onto other
> volumes. The procedure is basically, create the index, figure out which
> file in $PGDATA/base/oidofyourdbhere is the index, shut down postgresql,
> copy to file elsewhere, softlink it, restart apache. This setup will not
> survive reindexing or dropping / recreating the index.

It should also be noted that this hack doesn't cope well with tables or
indexes exceeding 1Gb. You'd have to separately symlink each segment,
and deal with creating new links every time the table/index grows past
another 1Gb boundary.

Tablespaces will be a lot nicer when we have them. For now, a RAID
array is definitely the best bet.

regards, tom lane

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message scott.marlowe 2004-04-09 17:20:43 Re: sequence value of the record just inserted.
Previous Message Josh Berkus 2004-04-09 16:02:00 Re: [PERFORM] Raw devices vs. Filesystems