How to Patch and Upgrade Your Databases at Scale

August 2, 2020 | 6 minute read
Timothy Mooney
Director, Product Marketing
Text Size 100%:

We just finished another webcast in the webcast series, Deep Dive Series on Enterprise Manager. These webcasts have been popular and rewarding to be a part of. Attendance has been strong, and interactions on Q&A have been active, in-depth, technical topics, evidence to the value of having experts online during the webinar.


Today’s session was anchored by Rene Fontcha, Master Principal Solutions Architect.  In short order he went through the challenges many of you face when patching and upgrading. It can be labor-intensive, human errors can slip in, and without standardization, the number of variables that can go wrong can be significant.


The next step:
Use automation, and Enterprise Manager (EM) can help. Many, if not most of you are already using Enterprise Manager with the vast majority of you leveraging Diagnostic and Tuning Packs for troubleshooting problems, and tuning database performance. What Rene discussed in depth is the next step for many DBA teams wanting to advance their skills by leveraging Enterprise Manager to automate regular tasks, including provisioning, patching, and upgrades and more.


Business Drivers

Rene focused on key business drivers that quickly translate to technology and process solutions. How do you deliver better performance, more efficiently, scale better, and move faster to meet business needs? His answer;

  • Consolidate databases on fewer systems
  • Standardize on fewer database configurations
  • Automate regular processes like provisioning, patching, upgrades and more

He’s seen how effective these steps can be with many companies who have applied these solutions.

Why now?

Rene spends a lot of time with IT organizations and DBAs at many companies and has heard the growing concerns for doing more and doing it with less, doing it at higher levels of quality, and increasingly, with more and more emphasis on security. As most of you know, security breaches are often the result from not staying current with security patches. 

What is your patch process?
Before we got into the details, Rene provided the incentive. He identified the progression or maturity if you will, for incorporating fleet maintenance into your processes. Of course, many use OPatch and write scripts to enhance the process.  Some leverage conventional processes using Enterprise Manager and the focus of today is using the integrated automation available as a part of Fleet Maintenance, a feature of the Database Lifecycle Management Pack. He’s seen the work involved to implement, and the payoff teams are getting is worth the effort when they institute the process.

Using 10 clusters as the standard payload from a real customer case, Rene shows the progress from 2 weeks to patch with OPatch and scripting, which goes down to 4 days with EM conventional patching, down to 4 hours when EM Fleet Maintenance was used.

A picture is worth 1,000 words.

Fleet Maintenance Model

There are 3 high-level steps to the process.

  1. First, standardize. Rene shared how to use Enterprise Manager advisors to reduce your configuration sprawl down to a minimum number of standard gold images.
  2. Second, subscribe databases to the gold images. Again, Enterprise Manager provides validation tools to ensure the subscription is valid.
  3. Third, is deploy. Here again, Enterprise Manager gives you some nice functionality. First, you apply the patch or update out-of-place, meaning in a new home. Once that work is completed, the only downtime remaining is switching the live environment to the new home. As you may hope, the Enterprise Manager toolset allows rollback, and clean up.

What’s nice about using Enterprise Manager Fleet Maintenance is it is designed for virtually every version and variation of the database, including RAC, Multitenant, and Exadata as well as Oracle database cloud services to mention a few and most of the regular database maintenance process from provisioning, patching, cloning, updates, and upgrades.


I don’t want to say it’s as simple as 1, 2, 3, but in summary, it starts with detecting configuration pollution.

The second step is creating gold images and subscribing databases to them.

The final step, in summary, is to deploy or propagate the gold images out-of-place, then switch the inactive updated homes with the active outdated homes.

After the overview, Rene went in-depth on what the process he would demonstrate later show in the EM console. He literally went through the whole process including a demo of rollback, and switching, and validating results were as expected.

I’ll post a recording of the demo as soon as I can here.

We also talked about a case study with Comcast. The discussion was around a large database estate where they have a large developer team who often had to wait months to get databases provisioned. After standardizing and implementing Fleet Maintenance, they were able to get provisioning down to minutes and actually used a self-service portal so their developers could manage their own destiny around deployment schedules. In the end, they had a much more responsive, predictable development environment that was patched automatically and saved maintenance downtime by using out-of-place patching.

Return on Investment

As you know, none of this comes without effort. I’ve from many customers who want to understand the cost and benefits before endeavoring on this journey.  It can be a difficult question to answer, there are so many facets involved, from people’s time, to license costs to equipment and more. To begin to answer that question, I want to point to a good paper written by Pique Solutions. They gathered detailed data from a good number of in-depth interviews with Enterprise Manager customers. They quantified the costs and benefits they saw from these customers and in addition to reporting on individual cases, they also normalized the results so other companies can apply some of the metrics to their business cases. It’s a well-written piece, and worth a read.

We have experts in the field who are versed in doing a consult on this very topic. If you are interested, contact your Oracle sales rep and let them know you are interested.

display on-demand recording

These are the resources Rene made available in his presentation.




Join the conversation  


Timothy Mooney

Director, Product Marketing

Experience across business development, product management, product marketing in computer hardware to cloud services.

Previous Post

Security Insights for your web apps with OMC Log Analytics

Nazih Bachir DJOUMI | 11 min read

Next Post

Oracle Management Cloud Release 1.50 Is Now Available!

Daniela Hansell | 2 min read