Re: Database practices/operations evaluations.

From: Michael Austin <>
Date: Mon, 06 Apr 2009 22:29:22 -0500
Message-ID: <qMzCl.11690$> wrote:
> All -
> I just realized that I did not provide sufficient information. We have
> one project where 12 DBAs are working 24*7 to support
> 250 databases. RAC and Replication are also in uses. Most of the
> databases are being used for 3 tier applications.

Hmmm. 20:1 - not bad odds... at my previous employer the ratio was almost 90:1 - don't try this as it is not for the faint of heart. 3 environments required 2-3 full-time DBA's 24x7.

Also more than 30+ RAC clusters and 2 data warehouses (both RAC'ed) that stored more data in one day than [most likely] all of your 250 db's combined. Critical databases used replication for DR purposes to a second and sometimes 3rd site.

> They got OLAP/OLTP and few DSS instance also. It is a 24*7 support for
> almost all the DBA tasks from installation to manage RAC/Replication
> and troubleshooting.
> I do not know what all other information I need to include here. I
> work for a diff. project and has been asked to evaluate
> this project based on best database practices and processes.
> If this were to work technically fix one or two problem I could have
> done this or I would have feel confident. But this type of
> task I have never done before. I know it required to be there for some
> time to understand the environment and implemented practices.
> I just thought to put this question for some suggestions who already
> had done this kind of job.
> Regards,
> Lyxx

The best thing I can say is that EACH database has it's own documentation on how it is configured, why it is configured that way, problem logs - and more important: how to fix it when it does <some weirdness>, how to fix it when it does standard errors ORA-600's can't connect, space issues etc... and lastly - ALL DBA's have access to ALL of this documentation 24x7.

If you want to reduce overhead and a lot of those after-hours calls about space issues - tell mgmnt to SPEND THE MONEY and get the space needed, not just the minimum space required. Use ASM with ASSM on your OLTP stuff, use ASM on your RAC clusters. If you have more than sufficient space AND using ASM, most of your late night "out of space" issues go away. It is HIGH time DBA's stopped fixing symptoms and started fixing the problems. Kludging your way by {re-}moving data, archive log files, etc... is just a symptom of the real problem - cheap management that won't spend the money to by the required storage to RUN their database. Received on Mon Apr 06 2009 - 22:29:22 CDT

Original text of this message