Re: Large Objects and and Vacuum - Mailing list pgsql-general

From Simon Windsor
Subject Re: Large Objects and and Vacuum
Date
Msg-id 003d01ccc955$b4726220$1d572660$@cornfield.me.uk
Whole thread Raw
In response to Re: Large Objects and and Vacuum  (John R Pierce <pierce@hogranch.com>)
Responses Re: Large Objects and and Vacuum
List pgsql-general
Hi

Thanks for the response.

I am new to small IT company that have recently migrated an Oracle based
system Postgres. The system stores full XML responses, ranging in size from
a few K to over 55MB, and a sub set of key XML fields are stored on a more
permanent basis.

The database design was thus determined by the previous Oracle/Java system,
with empty LOBS being created and data being streamed in.

The data only has to be kept for a few days, and generally the system is
performing well, but as stated in the email, regular use of vacuumlo, vacuum
and autovacuum leaves the OS disc space slowly shrinking.

As a last resort this week, I'm going to get 500+GB of extra file store
added, add a tablespace and move pg_largeobjects to this area. Then use
CLUSTER to rebuild pg_largeobjects back in the default tablespace. This
should fix things I hope, and if needed I'll use Cluster regularly.

Simon


-----Original Message-----
From: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org] On Behalf Of John R Pierce
Sent: 02 January 2012 11:18
To: pgsql-general@postgresql.org
Subject: Re: [GENERAL] Large Objects and and Vacuum

On 12/30/11 3:54 PM, Simon Windsor wrote:
> I am struggling with the volume and number of XML files a new
> application is storing.

how big are these XML files?  large_object was meant for storing very
large files, like videos, etc. multi-megabyte to gigabytes.   XML stuff
is typically a lot smaller than that.

me, I would be decomposing the XML in my application and storing the data in
proper relational tables, and only generate XML output if I absolutely had
to send it to another system beyond my control as its easily one of the most
inefficient methods of data representation out there.



--
john r pierce                            N 37, W 122
santa cruz ca                         mid-left coast


--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make
changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general


pgsql-general by date:

Previous
From: Jasen Betts
Date:
Subject: Re: [partition table] python fetchall or fetchone function can not get the returning rows
Next
From: Nulik Nol
Date:
Subject: asynchronous api questions