Kscope23 Recap

I’ve been back for a while but thought I’d post some plans for my Kscope content and talk about the conference itself now that it’s in the bag.

I arrived late Friday night and things were pretty quiet. I spent some time in the bar trying various bourbons and whiskeys before calling it an early night.

The next morning I ran into Connor McDonald (check out his amazing content on Youtube). He mentioned that he was planning to go visit the famous Tom Kyte (the original creator of Ask TOM, and author of one of the very best Oracle books, Expert Oracle Database Architecture) but apparently Tom fell off a ladder and broke some ribs. Hopefully, Tom will recover soon. A bit later I joined the team going to the community service day event. They gave us matching shirts to wear and we boarded buses to Mile High Behavioral Healthcare and did a bunch of cleaning, gardening, and painting. I was on the painting crew and painted a gazebo and a few tables with a great team of folks. I ended up buying MHBH some picnic tables after we went to pick up one of their existing picnic tables and it essentially disintegrated. When one of the guests of MHBH asked me “How do I get one of those shirts?” I gave him mine. While the shirt was cool I’m sure he’ll get more use out of it than I would have.

By Sunday the rest of the Insum team had started to show up and we had a great time re-connecting.

Both of my main presentations were on Monday. Thankfully I’d delivered both before, so I was really comfortable with the content although it was going to be my first time delivering my “You know you might have a bad data model when…” presentation in just thirty minutes (I’d been given one of the 30 minute sessions for this presentation). It’s a bunch of slides with the type of queries that you would either see in your applications or queries you can run against your application schemas to see if there might be opportunities for enhancements. Upon the advice of Steven Feuerstein, another member of the Insum team, instead of starting with the theory of normalization, I started with the actual queries. Since the theory portion would take about 5 minutes, my plan was to cut it off at 25 minutes and jump to the theory. I set an alarm on my phone for 25 minutes, let the audience know what I planned to do, and dove in. When I finished my queries section, I glanced at my watch to see how much time I had left. As I was glancing at my watch, the alarm on my phone went off! It was perfect and I got a bit of applause from the audience. I finished off the theory portion and then got a lot of good feedback including from some of the very experienced folks in the audience (Oracle ACE directors, etc.).

Later in the day, I did my “APEX and ORDS for DBAs and System Admins” presentation. While I’ve delivered the content at other events before, I always update my presentations to the latest and greatest software, and with the very frequent updates of both ORDS and APEX I had to update everything a few days before the conference.

This presentation is actually about 2 to 4 hours of content, but I only had an hour to deliver it. Basically, I cut the presentation in half and gave folks a preview of what would be in the 2nd half if I had more time. I also went through the first half of the presentation pretty quickly. The premise of the presentation is that people often come to Kscope and they see all these really cool APEX solutions, but then when they go back to their own IT department the DBA or System Admin just says “No.” The reason for the “No” can be anything they feel like coming up with at the time (“Security”, “Too complicated”, “I don’t know how to do it”, etc.) but the conference attendee doesn’t get the very cool APEX and/or ORDS features that they saw at the conference. To solve this problem, I broke the first half of the presentation into three sections.

1. Install the APEX database component and patch it in the database.
   This section shows that APEX is just a component of the database (it shows up in the dba_registry view) and, by showing what is happening in the database we see that there are NO security changes in the database when you do this. It also showed how the installation and patch take under 10 minutes. On my boxes, the APEX install takes usually takes under six minutes and the patch takes under 30 seconds.
2. Now that you have the APEX component in the database, you have access to bunches of very cool APIs. Some of those APIs enable you to reach out of the database to access things on remote servers if the DBAs allow it. I show exactly how to enable this and how to create and manage an Oracle TLS wallet (yes, people and Oracle often refer to this as an SSL wallet, but we really stopped using SSL back in the 1990s. It’s been TLS for over 20 years… and it really is a TLS wallet). Wallet management can be very tricky and I can’t tell you the number of times I’ve seen very bad wallet setups on Oracle servers. I explain the various wallet types (TDE wallet, SEPS wallet, XDB wallet, and TLS wallet) and show how I build and maintain them.
3. Finally, we get to the last step which is setting up and configuring ORDS for a production-ready deployment. While Kris Rice, the Oracle lead for the ORDS team, disagrees with me, I really don’t like deploying ORDS with /ords/ in the URL. As Sir Tim Berner’s Lee explains, cool URLs don’t have the technology used to deploy them in the URL. Yes, that link is from 1998. I figure when Kris is knighted by the King he can tell me I’m wrong. I also show how to maintain ORDS over time. I show how to create the best

At this point, I’ve covered the first half of the presentation but an hour has passed and I have to just show folks what the next three sections are:

• Adding a database resource manager plan that ‘just works’ for APEX deployments.
• Building a Systemd service to run ORDS.
• Deploying the ORDS service as a ‘regular’ operating system user instead of the root user. The root user can, of course, deploy ORDS on port 443 (the default HTTPS port) trivially, but some folks would prefer not to run ORDS as root, so I show how to do this.

My plan is to take all of that content and publish it here (update: The entire 2-hour APEX & ORDS for DBAs and System Admins is now live!), but it will take a while. Life is very busy and I’m not exactly sure when I’ll be able to finish it all. Until then, all of the code and examples from my both presentations can be found in the following Github repos:

• you-know-you-might-have-a-bad-data-model-when
• APEXandORDSforDBASandSysadmins

Finally, I was a co-presenter with Cary Milsap of Method R in a presentation on tracing APEX. It was also well-received and sparked a lot of interesting discussions.

When I do get the time to post everything from my presentations I’ll try to remember to come back and here and link it in. Update: Here’s the link to the whole presentation from start to finish!

Until then, happy DBAing, System Admining, APEXing, and ORDSing.