RE: Database Outages - Best Practices

>> Table reorgs (Typically after huge deletes as in a table purge..)

Niall Litchfield responds:
"I'm not necessarily sure I agree with this one either, it maybe that the cost of the downtime doesn't justify it."

So you purge data, drop a table from 27G to 16G , you're running short on disk space and reclaim that disk, performance is OK before the outage.

Reorg a table and get rid of the 'swiss cheese' recover 9G of disk space (THIS is in just one table) improve performance by dropping highwater mark and compressing the extents getting rid of the holes that need to scanned across and you don't think it is worth a few hours of the database being unavailable once a year.

Users and management are tickled over the performance improvements gained by this step. Guess that's all the justification needed.

-----Original Message-----
From: Niall Litchfield [mailto:niall.litchfield_at_gmail.com]=20 Sent: Thursday, February 17, 2005 3:05 AM To: Hollis, Les
Cc: stephenbooth.uk_at_gmail.com; ryan_gaffuri_at_comcast.net; mfontana_at_verio.net; oracle-l_at_freelists.org Subject: Re: Database Outages - Best Practices

On Mon, 14 Feb 2005 15:26:54 -0600, Hollis, Les <Les.Hollis_at_ps.net> wrote:
> We call it a 'maintenance window'...

>=20
> ALL databases require periodic maintenance

This I agree with

> Rebuilding indexes

This I don't - but if you are going to do it - do it online without an outage.=20

> Running in new application code
> Table reorgs (Typically after huge deletes as in a table purge..)

I'm not necessarily sure I agree with this one either, it maybe that the cost of the downtime doesn't justify it.

> Tablespace rebuilds

As before. What is the cost of not doing it compared to doing it. In dollars.=20

I'm interested that you don't include patching the database or os in your list.=20

--=20
Niall Litchfield
Oracle DBA
http://www.niall.litchfield.dial.pipex.com

--
http://www.freelists.org/webpage/oracle-l