Re: Backing up large databases

From: "Uwe C(dot) Schroeder" <uwe(at)oss4u(dot)com>
To: pgsql-admin(at)postgresql(dot)org
Cc: Steve Burrows <steve(at)jla(dot)com>
Subject: Re: Backing up large databases
Date: 2006-04-29 03:20:56
Message-ID: 200604282020.56811.uwe@oss4u.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Why don't you set up a second server with identical database and use Slony to
replicate the primary one there. You can then do whatever you want on the
mirror - dump it on a tape, etc.

On Friday 28 April 2006 08:57, Steve Burrows wrote:
> I am struggling to find an acceptable way of backing up a PostgreSQL 7.4
> database.
>
> The database is quite large, currently it occupies about 180GB, divided
> into two elements, a set of active tables and a set of archive tables
> which are only used for insertions.
>
> I ran pg_dump -Fc recently, it took 23.5 hours to run, and output a
> single file of 126GB. Obviously as the database continues to grow it
> will soon be so large that it cannot be pg_dumped within a day. Running
> rsync to do a complete fresh copy of the pgsql file structure took 4
> hours, but later that day running another iteration of rsync (which
> should have only copied changed files) took 3 hours, and I cannot afford
> to have the db down that long.
>
> Anybody with any ideas? The database is being used as the backend for a
> mail server, so it has transactions 24 hours a day but is quieter at
> night. I want to be able to back it up or replicate it on a daily basis
> with minimum downtime so that the mail backlog doesn't get too large.
> Ideally I want the first generation of backup/replica going onto the
> same machine as the original because the volume of data is such that any
> attempt at network or tape backup of the live files will require too
> much downtime, once I've got a backup then I can copy that out to other
> NAS or tape at leisure.
>
> If anyone has experience of safeguarding a similarly large PostgreSQL
> database with minimal downtime I'd be delighted to hear.. The machine
> is running 2 Xeons, 4GB ram and a half-terabyte RAID10 array on a DELL
> PERC scsi subsystem, with a load average of around 0.5 - 0.6, so it's
> not exactly overstretched.
>
> Thanks,
>
> Steve

--
UC

--
Open Source Solutions 4U, LLC 1618 Kelly St
Phone: +1 707 568 3056 Santa Rosa, CA 95401
Cell: +1 650 302 2405 United States
Fax: +1 707 568 6416

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Warren Little 2006-04-29 12:52:55 Re: questions on toast tables
Previous Message Jim Nasby 2006-04-29 01:34:15 Re: WAL recovery question - 0000001.history