Tuesday, April 9, 2019

COLLABORATE19: Day 2

9:15 AM–11:30 AM - Hands on Lab: Migrating WebLogic Applications to Kubernetes in Oracle Cloud - @richardintx & Thomas Palzkill

Picking up where I left off the day before, I'm back in the same room, with the same Oracle experts, back at https://go.oracle.com/hol but this time ready to migrate WebLogic to Kubernetes! Exciting! Except.....when you see most of the introduction of the lab is the same as the day before, yet it does not work at the end this time, and the solution is to delete the Kubernetes cluster. That part is not even the bummer, the bummer is when it is giving the whole class an error message about APIs taking longer and just stalling us all for an hour until the instructors find out the Management backplane is having an issue, and to resolve it we have to switch to another Oracle Data Center. That is.....not really reassuring about going to the Oracle Cloud. Sure, everything in their DC is still running what was running (per Richard) but if the 20-30 people that might have been hitting the Management part of that DC helped to overwhelm it in some way, that seems like a bad thing. When a Cloud is not working, is it a StormCloud then?

Lunch! I finally got to have lunch when everybody else was eating! It was a novel experience! :}

12:45 PM–1:45 PM - Designing Scalable Oracle Database Solutions in AWS - Nick Walter of @houseofbrick

The one big takeaway here from Nick, is that the licensing costs of software going to AWS will be the biggest costs you will have and any AWS options are peanuts comparatively. Also, I loved his quote "a well tuned database is always limited by I/O" which is in relation to how to buy I/O options from AWS and that is good to know/remember! More importantly, instance local storage is fast and dangerous because it does NOT survive instance reboot so you CANNOT put Oracle files there!!

2:00 PM–3:00 PM - An AWS DMS Replication Journey from Oracle to Aurora MySQL - @MarisDBA

Maris' presentation was more of an overview from one perspective, but since I have never touched any of the AWS, DMS, or Aurora MySQL technologies he talked about I certainly did not mind at all! I thought it was smart to start us off on some "basics" going over the motivation (DB contract expiration, company wanted to be cloud native), landscape (6 month hard deadline, multiple schemas, cannot move all apps at the same time, too much risk), and complications (cross schema FKs, other apps needed to have access to the main schema that was being moved). The rest of the presentation went by pretty quick, with the ideas and solutions which were put forward, and eventually rejected, until they came upon a solution which solved all their needs and addressed the complications of the project. There was a lot which Maris went through to get the solution implemented, and I am looking forward to seeing if maybe we can replicate some of what he did in order to get something small over to AWS.

3:15 PM–4:15 PM - Cloud 2020 Where We're Going from Here - Various Panel

Going into this, with only 5-10 sessions available for the entire conference, I knew that the likelihood of this being a sales pitch were high, but it did not start off like that! The main concern about moving to the cloud is not really about infrastructure costs, it is more about people (retire old skills, enable new skills), security (60-85% of all security breaches were preventable), and capabilities. Then going to the next presenter on cloud best practices, it is more a question of WHEN we will go to the cloud, not really a question of IF. Most of their presentation was about what services their company offers, which is not really where the cloud is going, even if they did suggest reading the book Leading Change by John P. Kotter. The last presenter on deck started talking about their cloud conversion story, focused on JD Edwards, and after about five minutes of talking about basic project management and app specific details, I decided to leave the session. This was unexpected as the small room was completely full, and LOTS of people were leaving, likely due to the same reason I did, and the abstract really led us to believe it would go differently.

4:30 PM–5:30 PM - 10 things you should know about Oracle and Kubernetes - @MNEMONIC01

I will not list all of these (mostly because I missed a few along the way) but there were a couple that stood out like: Oracle adopts Kubernetes, you can do it yourself even using Google packages on RedHat Linux if you are not on Oracle Linux, you should look into Helm & Tiller which is a release and package manager, there is a WebLogic migration to Kubernetes tool coming soon which will even migrate 11g.

1 comment:

  1. Big data modernization solutions should understand the need of Data, and they should work to build more appropriate services to meet the requirements of their clients.

    ReplyDelete