pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Jim Nasby <jnasby@pervasive.com>
To: Benjamin Krajmalnik <kraj@illumen.com>
Cc: jd@commandprompt.com
Cc: pgsql-admin@postgresql.org
Subject: Re: File system backup question
Date: Tue, 8 Aug 2006 11:48:56 -0500
Message-ID: <E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com> (raw)
In-Reply-To: <BF337097BDD9D849A2F4B818DDB27987252585@stash.stackdump.local>
References: <BF337097BDD9D849A2F4B818DDB27987252585@stash.stackdump.local>

If you must use inserts, you'll get much better performance if you  
wrap the restore in a transaction. I think someone came up with a  
clever way to do this on the command line, or you could edit the dump  
file. Note that if you do this and there's any errors, the restore  
will fail.

Also, greatly increasing maintenance_work_mem during a restore can  
help with the index builds.

On Aug 3, 2006, at 11:05 PM, Benjamin Krajmalnik wrote:
> I believe I have to use onserts due to the problem with the records  
> with the improper utf encoding.  According to the docs this will  
> cause the restore to fail.
>
> I will try it as well as a file system restore.
>
> -----Original Message-----
> From: Joshua D. Drake <jd@commandprompt.com>
> To: Benjamin Krajmalnik <kraj@illumen.com>
> CC: pgsql-admin@postgresql.org <pgsql-admin@postgresql.org>
> Sent: Thu Aug 03 21:01:54 2006
> Subject: Re: [ADMIN] File system backup question
>
> Benjamin Krajmalnik wrote:
> > I am currently running PG 8.1.4 windows.
> > The system is for a real time monitoring application (so downtime  
> needs
> > to be minimized if at all possible).
> >
> > I will be migratng to a new server running PG 8.1.4 on FreeBSD 6.1.
> > I have been playing with various options for migrating the data.   
> The
> > database is currently about 8.5 GB.
> > pg_dump took about 90 minutes.
> > pg_restore has been running for over an hour and is abou 4% done  
> (based
> > on row counts on the tables).
> >
> > I used the pg_dump option to insert records (as oposed to use the  
> copy
> > command - I don't know which would be faster for the restore).
>
> That is why it is taking so long.
>
> Do not use inserts
> Turn off fsync
> Crank up shared buffers
> make sure you have lots of checkpoint_segments.
>
> Joshua D. Drake
>
>
> >
> > In any case, I was thinking about performing a file system backup.
> > My concern is that I am movinf between different OS's (although  
> both are
> > i386 architecture).  I will probably try it tomorrow to see if it  
> works,
> > but was just wondering if there are any caveats which I should be  
> aware
> > of.
> > I hope this works, since for my scenario, this will have the least
> > downtime.
> >
> > Thanks in advance for any advice.
> >
> > Regards,
> >
> > Benjamin
> >
> >
> >
> > PS.  For those who have asked about performance, Windows is much  
> slower
> > than FreeBSD.
> >
> > Machine specs are as follows:
> >
> > Windows 2003 server, HP DL360, SCSI RAID-1, 2GB RAM, 2.8GHz XEON HT
> > FreeBSD 6.1, SuperMicro 5014C-T, SATA RAID-1, 1GB RAM, P4 3GHz
> >
> > My stored procedure call went from 47ms avg to 6 ms average  
> execution
> > time on server, as reported by turning logging on
> > All queries in general appeared to run between 5x to 10x faster.
> >
> >
> >
> >
> >
>
>
> --
>
>     === The PostgreSQL Company: Command Prompt, Inc. ===
> Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
>     Providing the most comprehensive  PostgreSQL solutions since 1997
>               http://www.commandprompt.com/
>
>
>

--
Jim C. Nasby, Sr. Engineering Consultant      jnasby@pervasive.com
Pervasive Software      http://pervasive.com    work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf       cell: 512-569-9461



view thread (4+ messages)

Message-ID: <E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com>
Permalink:  ../E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com/
Also on:    postgresql.org/message-id/E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-admin@postgresql.org
  Cc: jnasby@pervasive.com, kraj@illumen.com, jd@commandprompt.com
  Subject: Re: File system backup question
  In-Reply-To: <E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com>
  References: <BF337097BDD9D849A2F4B818DDB27987252585@stash.stackdump.local> <E69BB30F-F365-4E5D-B573-FCECF554B96A@pervasive.com>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox