This is the first post in a new blog series by the IBM App Platform SWAT team. Our team engages worldwide in critical situations for IBM WebSphere Application Server (WAS) (traditional and Liberty) with deep expertise in WAS, Java, operating systems, networking and other technologies to drive client success. This blog series will focus on practical lessons from the field that you can use to improve the health of your systems.
Today, we'll cover the free lab we've created to help you learn how to troubleshoot and performance tune WAS deployments: the WebSphere Application Server Troubleshooting and Performance Lab on Docker
. Earlier this month, we refreshed the software in this lab to include all the latest and greatest versions including Liberty 22.214.171.124. For those already familiar with this lab, stay tuned for future blog posts that dive into fresh technical topics such as tips for WAS on OpenShift, theoretical topics around performance tuning, deep dive topics like tuning TCP/IP, tooling, and much more.
The WAS Troubleshooting and Performance lab is really use to use: 1) Install Docker Desktop
on your computer, 2) Download and run the lab
, and 3) Start a VNC or remote desktop session with the password websphere
to run the lab. Everything is fully self-contained so you can even run the lab without internet. If you don't have permissions to install Docker Desktop, we recommend setting up a Docker environment on Linux somewhere in your company, run the lab there, and then remote into it. If you'd like to see a video presentation of the lab and a demo showing how to run and use it, check out a replay of a talk
we gave on it last year.
Another nice feature of the lab is that you can either run through it step-by-step or jump right into a particular topic. It's been designed so that all the pre-requisites are baked in and you can start anywhere you'd like if you're only interested in learning a particular type of problem or how to use a particular tool. We based the content on the things we see in the field that affect the most customers. You can see all the topics in the Table of Contents in the lab document
, including CPU analysis, thread dump analysis, Java garbage collection analysis, Java memory analysis, and more.
Once you've remoted in, you'll be at a Linux Fedora 33 GUI session. Then, simply follow the instructions in the lab document
to start simulated web traffic with Apache JMeter pointed at either the pre-installed WAS traditional or WAS Liberty servers and investigate issues with pre-installed tools such as GCMV, TMDA, MAT, HealthCenter, and more. As background and for ideas of what to do in addition to the self-paced lab, you might also find interest in these other recent presentations we gave: