Last Saturday (Aug, 5th) it was held in São Paulo the 8th edition of GUOB Techday. This event is one of the most important event for DBAs and Oracle Professionals, it is part of the Latin America Oracle Technology Network Tour, which brings top professionals of different Oracle technologies to speak and exchange ideas with the Oracle community.
I will speak about the event in general and my experience throughout the day.
Arrival / Check-in
I arrived around 8:15 AM at Uninove Vergueiro, the university where the event was being held. The organization was magnificent and I could get my credentials ready in no time. I then went ahead to coffee and met already with a lot of colleagues, friends and speakers. The exchange of ideas and chatting was great while enjoying a whole lot of awesome food and beverages. The networking coffee during the day were pretty similar to this, so I will not cover each interval between sessions, they were all awesome.
At 8:30 I got a seat as the opening started. Eduardo Hahn thanked Uninove for providing the infrastructure for the event as this was the first time on a university instead of a hotel, which is nice as this was due ever growing number of participants and concurrent sessions. After a brief overview of the past and future of GUOB, Pablo Cicarello took the stage to talk about the Oracle ACE Program and encourage Oracle professionals to engage and participate more and more in the community. Eduardo Hahn took the stage back for a few moments to say a few more words and call the first speaker of the day: Mike Dietrich (Master Product Manager Database Upgrades & Migrations @ Oracle)
Upgrade to 12.2 – Live and Uncensored – Mike Dietrich
This was the first session of the day, everybody participated in this one as it was held in the main room in the first floor of the university. In this session, Mike introduced us to a few cases of success for Oracle Upgrades, some challenging scenarios such as moving from 8i to 12c and performed 2 upgrades live, one from command line and one from DBUA (graphical interface). I’m suspicious to speak about Mike’s presentation since I always watch them whenever possible. I’ll try to summarize a few interesting things that I believe to be of great help: Do not wait too long to upgrade, you will just trap yourself and make the whole process a lot harder and painful. Use the command line, there are awesome features there like parallelism and resume option. Finally, download the virtual machine to practice the upgrades and test it out from his website: https://mikedietrichde.com/hands-on-lab/
Ensure Performance Stability when upgrading Oracle Database – Mike Dietrich
The second session I attended was also from Mike, the sessions now were moved up to the 8th floor of the university. In this session Mike approached some key factors to ensure a successful upgrade. It is not rare the cases where after upgrading a database results in users complaining about how their systems were faster before, so be prepared! Take a baseline snapshot of your workload data before the upgrade, patch your environment, take care with optimizer parameters settings and most important: TEST! TEST! And… TEST! Cool tip about tools to allow a great testing scenario: Since Enterprise Edition doesn’t come with SPA (SQL Performance Analyzer) license included, it is possible to setup it in the Oracle Cloud, as in the cloud the license is included. You can now capture your SQL into SQL Tuning Sets and run them against the SQL Performance Analyzer in the cloud, anticipate issues and fix them, shutdown the cloud database (EE High Performance or Extreme Performance) and move on with your on-premises upgrade.
Database Security with Transparent Data Encryption – Adriano Bonacin
Following session presented by Adriano Bonacin (Pagseguro UOL), demonstrated from the concepts of cryptography to details of the process involved in symmetric and asymmetric encryption, hashing algorithms, what is salt and what it is used for. Moving on to the practical aspects of the TDE, he demonstrated the techniques and encryption types by using column encryption vs tablespace encryption, how the different types of encryption might affect storage utilization. He presented also the procedures to implement TDE and administer it, this included the creation and maintenance of the wallet, how to change their keys and how to set it up to auto open, so we can access the data encrypted after a database bounce.
Maximizing Oracle Cloud Buffer Cache Throughput – Craig Shallahamer
Craig Shallahamer from OraPub presented this very interactive session on how to optimize buffer cache throughput, even though the title of the session said “cloud”, the tips and techniques presented by him can be applied to every Oracle Database, both on-premises and in the cloud. He explained the internals about the buffer cache when we see “free buffer” wait event and what is going on in the memory and how that affects the performance. A lot was said from when the database needs to do an physical I/O to how Oracle keeps or eliminates the blocks from memory using the MRU and LRU lists.
Getting the most out of Oracle Grid Infrastructure – Franky Weber
This session presented by my friend Franky Weber (Pagseguro UOL) is really a thunderstorm of information about Grid Infrastructure with great emphasis on new features of Oracle 12c. In this presentation, Franky introduces us to good practices on maintaining ASM with ASM Filter Driver, setup of diskgroups and a ton of information on Flex ASM, including new features such as ASM File groups and Quota groups. Practical examples can be obtained from his blog at http://loredata.com.br/blog/, including the setup and configuration of ASMFD, Diskgroups with Flex redundancy and much more. During the session he also presented a few parameters worth changing from the default values, to ensure a better response time on failures from nodes in a RAC configuration, what is and how to move GIMR to a different diskgroup. One of my favorite things in his presentation is the “explain work for” feature for ASM operations.
How to diagnose Random Oracle Cloud Performance Incidents using ASH – Craig Shallahamer
And we finally arrive to the last session of GUOB Techday 2017. Craig Shallahamer took over the main stage to demonstrate how we can use ASH to drill down on active session history and find the needle in a haystack. He used the bloodhound toolkit to find a particular case of a deadlock impacting a batch processing application. He took a step further to generate a script dynamically to be consumed by R and generate the visual representation of that deadlock.
Closing and conclusion
A couple of lucky fellows won tablets on a raffle help by Eduardo Hahn with the speakers, where the event finished and can be considered a huge success. As always it was great to be in the presence of such great professionals and have the opportunity to talk to them and learn from them.
I hope to see you there next year!