Introduction

Thank you very much for buying the second edition of Expert Oracle Exadata. Us current authors have been standing on the shoulders of giants while putting this together. Whenever writing a second edition of a successful book, the authors face the pressure of creating at least as good, if not better, edition than the first edition was. And good it was, the first edition. We hope that we have been able to provide you, dear reader, with a suitable introduction to Exadata. In fact, our hope is to give you enough information to get started with Exadata. It is not uncommon to find database administrators in situations where they have been introduced to Exadata. only to ask the question, “Now what?” We have tried to structure the book to help you answer this question. You will read about what Exadata is before diving into the various optimizations that make it so unique in the world of Oracle database processing hardware. While some of the material, particularly in the earlier chapters, paints a broad picture, we gradually go into a lot of detail. Access to an Exadata development system can help you a lot in understanding the more advanced material. We have tried very hard to make it possible for you to follow along, but please bear in mind that the Exadata platform is not static at all; new releases in hardware and software can change the documented outcome of commands and SQL statements. We will try to address major differences on our web site, http://www.expertoracleexadata.com/ and our personal blogs listed in Appendix B.

Note that we have used various undocumented underscore parameters and features to demonstrate how various pieces of the software work. Do not take this as a recommended approach for managing a production system! In fact, there is usually no reason to deviate from the defaults. Setting underscore parameters is allowed only with the explicit blessings from Oracle Support and as the result of a recommendation as part of a service request you raised. Remember that we have had access to a number of systems that we could tear apart with little worry about the consequences that resulted from our actions. This gave us a huge advantage in our investigations on how Exadata works across various hardware generations.

The Intended Audience

This book assumes that you are already familiar with Oracle. We do not go into a lot of detail explaining how Oracle works except as it relates to the Exadata platform. This means that we have made some assumptions about the readers’ knowledge. We do not assume, for instance, that you are an expert in Oracle performance tuning, but we expect that you are proficient writing SQL statements and have a good understanding of the Oracle architecture. Since Exadata is a hardware and software platform, you will inevitably see references to Linux administration in some of the chapters more closely related to the hardware. Do not be intimidated—as an Exadata administrator, there are only a handful of commands that you need to know in day-to-day managing of the platform.

A Moving Target

We had this exact same section in the introduction of the first edition of this book, and the message is still the same, even after all these years. What keeps us amazed to this day is the pace of development of the Exadata platform. It is not only hardware that evolves and keeps up with the development of new technologies, but also the software that is constantly pushing the limits of what is possible. A new software release does not require you to upgrade the hardware. Except for the very first Exadata system, the current Exadata software version is compatible with every hardware generation.

The changes mentioned in the previous paragraph include substantial additions of new functionality, visible in Appendix A in the Exadata Database Machine System Overview. As you can imagine, trying to keep track of what Oracle released at a rapid pace was the most difficult part of the project. Every chapter had to go through multiple revisions when new hardware and software was released. The latest version we try to cover in this book is Oracle 12.1.0.2.2 RDBMS with cell software 12.1.2.1.x. Unlike the first edition of this book, which came out when Oracle 11.2.0.2 was current, there are quite a few releases now that Exadata supports technically. From an Oracle Support point of view, right now you should probably be in a migration phase to Oracle 12c. This is one of the reasons we gave the latest RDBMS release so much space in the book, even though many users are yet to migrate to it. Another consideration while writing this book was that we had to be quite careful to cite the correct version when a new feature was introduced. If you only have just started with Exadata, you might find the release numbers confusing; however, once you have your first few weeks of Exadata administration under your belt, you will find that quoting Exadata cell software releases becomes second nature.

The way Exadata evolves will undoubtedly make some of the book’s contents obsolete, so if you observe differences between what is covered in this book and what you see it is probably due to version differences. Nevertheless, we welcome your feedback and will address any inconsistencies that you find.

Many Thanks to Everyone Who Helped!

We have had a great deal of support from a number of people on this project. Having our official technical reviewer take on writing a few chapters is almost an occurrence of history repeating itself. Writers and reviewers swapped roles to reply to the question, “Quis custodiet ipsos custodes?” We are also very grateful for everyone at Oracle who may have even known us from the first edition of this book and helped us overcome the stumbling blocks along the way. Finally, we want to give a big “thank you” to everyone at @Enkitec who helped keep the machines up and running, patched when a new release came out, and troubleshot when something seemed broken. The list of people is really long, so we won’t be able to mention everyone by name. However, it is fair to say that if you worked at @Enkitec while this book was being written, you almost certainly contributed—thank you.

The first book helped generate interest in the second edition, and we have published some research that was too comprehensive on our personal blogs and web sites, prompting e-mail, twitter, and comments to start flying our way once an article went online. The same is true for the feedback we had with the Alpha Programme; without the community’s feedback, this book would probably be less complete, and we would like to explicitly thank you for your comments.

And last, but not least, we would like to give a very special “thank you” to the authors of the first edition of the book, who allowed us to update what they wrote. Kerry, Tanel, and Randy have been instrumental in understanding the intended message of the chapters as well as chapter layout and tests. Without you, we wouldn’t have been able to finish the chapters while maintaining the spirit of the first book.

Who Wrote What

Following the tradition set in the first edition, we would like to list which of us worked on each chapter. The authors of the second edition (in alphabetical order) are Karl Arao, Martin Bach, Andy Colvin, and Frits Hoogland. It really was a team effort between all of us involved, and we cannot even think about counting the hours of useful conversations and instant messages exchanged among all of us to bounce off ideas and make sure that we did not overlap contents in our chapters.

• Karl: contributions to Chapters 5, 6, 7, 12
• Andy: Chapters 1, 8, 9, 14, 15, 16, Appendix D
• Martin: Chapters 2, 3, 5, 10, 11, 12, 13, 17, Appendices A, B, C
• Frits: Chapters 4, 6, 7

Have Fun!

Writing the book was, for the most part, fun for all of us—especially when we knew about a complex problem, but had trouble reproducing a situation allowing us to research it. The moment the experiment came to a successful conclusion, the moment when we had all the output and steps to reproduce it recorded in our log files, was very often a moment of great joy and also relief. We hope his book provides a platform from which you can build your own knowledge. Although having spent a lot of time with both Exadata and Oracle Database 12c, there are still things we learn every day. Somehow it still feels we are only scratching the surface, still.