Expert Session:- Kubernetes Industry Use-cases
On 8 March 2021 we at LinuxWorld had an amazing expert session with our experts as follows:
♦Mr. Neeraj Bhatt
Senior Technical Engineer- OpenShift, Redhat
♦Mr. Vijit Kuntal
Consultant, Infosys Belgium
♦Mr. Rushil Sharma
Customer Engineer Hybrid Cloud, Google
The first use-case presented by Mr. Vijit Kuntal was about how he helped a company to migrate from CVS to GIT. The company had around 1600 modules existing on CVS which were to be shifted to GIT.
To do so there are two ways
- git cvsimport: which is a git native tool to help us with this task
- cvs2git: which is a very powerful tool to do the task. Interestingly this tool or method was created in just about 45 days.
CVS is a Centralized Version Control System and it is used to store code in one centralized version. Git is a distributed version system where multiple developers can commit and it has also branches where we can merge, create, delete, but in CVS only the one person can commit or make changes in the central repository.
After this a CICD pipeline was created using Jenkins for end to end production system.
Our second use was delivered by Mr. Rushil Sharma, about how to he optimized a banks resources and saved them “thousands of dollars per day”
The system design and architecture had Global Traffic Manager(GTM) is designed to provide high availability by managing the public traffic and is based on OSI layer seven which is usually known as Application Load Balancer.
Chaos testing is one kind of testing that is done to test the performance of cluster. Various companies like Netflix use chaos testing in there infra as well.
Engineering cluster is where we deploy the code which is not complete just for testing purpose. SIT(system integration environment) is where we deploy code when it’s almost ready for production.
CPU requests refers to a request made for using CPU resource by application and CPU Limits are the maximum number of CPU cores that production can provide to the container. CPU leaks is a condition that leads to taking more and more CPUs from the physical devices. This can be monitered by tools like Prometheus & Grafana.
By reducing the CPU requests and CPU leaking the resource required to run application were reduced to from ~600 cores to ~180 cores and from that ~900 pods were launched whereas before due to 1 core per process there were only ~600 pods running.
Re-Purpose means using the same resource again and again as it gets free from other processes.
The third use case was delivered by Mr. Neeraj Bhatt, where he explained a security use case. As as solution he implemented SCC.
OpenShift provides some security context constraints (SCC) that control the action and permissions for a pod using CLI
Thank you so much Vimal Daga sir, Preeti Chandak ma’am and the entire LinuxWorld India team for giving us opportunity to learn from these industry professionals.
