Note: For the sessions I’ll embed the videos once they are released. Similar to the general session.
General Session
Big announcements
- Acquisition of Pivotal and Carbon Black
- Project Tenzu with Project Pacific bringing native Kubernetes into the vSphere space running Kubernetes on ESXi and stating that it can run over 8% faster than bare metal and 30% faster than a linux VM.
- Some other announcements but my mind was already blown.
Build Your First Kubernetes Cluster on vSphere with Kubeadm [CODE1151UR]
Kenny Coleman @kendrickcoleman
Packed house makes it hard to type, but obviously. Kubernetes is a main interest point for all of the attendees of VMware. Also gotta shout out to VMware Code as they booths are pretty slick and awesome at the same time.
- Kubeadm – De-facto under the covers use of creating your Kubernetes cluster.
- Checkout the Kuberenetes the hard way in git
- what is your distribution server(what OS)
- make your distribution configurable as a template.
- Important to find the pipeline to build it.
- Now time to look through all the ways to install Kubernetes.
- Networking! How do you inter-pod networking setup?(calicok, canal, kube-router,romana)
- Are you ready to install? Nope gotta figure out storage. Persistent storage is a needed process that needs to be done. Especially moving that storage to CSI(container storage interface)
- blah.cloud to setup the install. Now you can install Kubernetes.
- Resource pool with no resources in vSphere??? Use Clustername/resources in the file to utilize the cluster.
- Kubeadm runs upgrades for you for clusters. Documentation on what can and cant be done and the pre-reqs
- Project Tanzu still part of the whole solution.
- kubeadm normally runs with one master, now in Beta it can look into multi-master setup.
- KubeADM is in alpha for Windows in 1.16
- How to get started – https://kubernetes.academy/
- Go check out blah.cloud.
- Free HOL on PKS – goto: https://labs.hol.vmware.com/HOL/catalogs/catalog/1212 and search PKS
Cloud Native Fundamentals: Containers and Kubernetes 101 for the VI Admin [KUB1126BU]
Nathan Ness @nvpnathan & Scottie Ray @H2Only
- Basically starting off right where the previous session talking about all the components of Kubernetes
- But it all starts on a dockerfile(if running on Docker)
- One beauty of Kubernetes and Containers in general, is that once you have the template, file, w/e to setup your app. it can be run on whatever DLL you have.
- Kubernetes terms
- Node = Conainer host w/agent called “kubelet”
- Application Deployment File = configuration file of desired state
- Container Image = runs in a pod
- Replicas = copies of pods that must be running
- Moving from a Container to a pod, a Pod is the concept of a side-car idea as a container is one microservice whereas the pod has multiple microservices.
- Labels help, do that….
- Type:LoadBalancer is Key to the Kubernetes service as it takes the application and turns it into HA with its replicas.
A lot of Q and A time after the session. It seems like a lot of people are looking for Kunernetes answers to questions they haven’t really defined. There are a lot of great training out there for Kubernetes, but I’m definitely going to write up some stuff to go over and help define those answers. Gotta get home though… and remember to do it while renovating our house to sell..
Continuous Application Delivery with Code Stream: A customer journey [HBO1080BU]
With Juan Hernandez, and Deepak Kallakuri
[No Link]
- Trying to get code into a CI/CD pipeline is the hardest part of most companies. Figuring out how to create this process is paramount to keeping the process running well.
- Two different pipelines sets, one for infrastructure as code utilizing Terraform, Cloud Formation, Cloud Foundations, CAS, ETC. and Application Pipelines pushing code to an app(Using azure devops, Jenkins, spinnaker, etc.)
- Code Stream is integrated with Cloud Assembly in CAS so it makes sense that the combination of vRA and CAS flows.
- Northern Grumman went to 250 feature branch merges a day. That’s awesome.
- CD pipeline for code stream
Once I get the video up or this its worth a watch for the demo itself. The co-joining between CAS and vRealize is a bit easier to understand when they talk through it. Though I’ll admit I’m not a fan at what Code Stream looks like. Perhaps the gui, and I need to dig in more with the code.
Building and Controlling Modern REST APIs with OpenAPI and Kong [CODE3326U]
With Dan Illson @dan_illson http://www.cloudjourney.io
- Kong is a usable API gateway other things like ambassador
- API = application programming interface
- APIs bring multiple machines into one, as previously engineers didn’t care where the APIs were because normally they were all on the same machine, now they are all over the place.
- APIs have built ins like get and post to utilize information
- CRUD = Create, Read, Update, Delete
- Lets talk about REST API
- Key unit of measure of REST is a resource
- REST is a good process of push pull information from one place to another.
- OpenAPI Specification(Swagger 3.0)
- OpenAPI definition is created via a Yaml file.
- What is an API gateway? An element which coordinates and orchestrates how requests are processed.
- It routes requests to services
- Most commonly used as “front end” for sets of services/microservices
- May also be “internally” between the components of an application
- Kong – A popular open source API gateway implementation
- Written on nginx
- API driven configuration
- Modular via plugins
The Genius of Kubernetes [CODE1172UR]
Richard Lander
- Ingredients of Success
- API Driven System
- Kubernetes API is THE interface!
- Control Loops
- Reconciling State
- Kubernetes Control Plane
- Extensibility – Stacking building blocks
- Community – Compounding the success
- The Benefits of Kubernetes – Boost Productivity, Improve Reliability
- Links
Automating Builds and Deployments, aka CI/CD for Dummies [CODE3330U]
Tim Davis @vtimd http://www.cloudjourney.io
- This is focusing on Software development and what those tools are like
- Why automate
- You can know whats out there but you need to standardize and create policy around it.
- Also the manual process is bad… very bad…
- What is CI/CD
- CI = Continuous Integration
- Your Code base when updated is deployed and set into a repo, and pushed into production via CD
- CD = Continuous Deployment(Delivery) – Sending that code repo into production
- Taking the code thats been built and sent into the production servers for actual use.
- CI = Continuous Integration
- How does it work?(Demo)
- Basically starting with a GitLab repository(because not a fan of GitHub).
- Once the code is local, you can run the docker build and then push it to the docker repository(docker.io)
- The build yml file indicates stages such as Scanning for claire scanning to run and find CDEs on any files they are using.
- Performance scanning with Locust performance to check the application
- One thing is a variable that can be piped into the yaml via the secrets stored in GitLab
- How do you get all this into the pipeline? Automate one piece at a time, one stage at a time. Automation the build stage, if that works Automate the next stage etc. until you have a good and repeatable CI/CD pipeline.
- Make it up as we go along.
Things that happened today!
vBeards with my VMug Leader!! I had to faux it… 😦
#vBeardsRockSFO @wmichel and @vNerdyNate pic.twitter.com/dCyc7iw4U8
— Wences Michel (@wmichel) August 27, 2019
I found CODY!!
Found @Codydearkland!!! He looks like he gives the best hugs… pic.twitter.com/7RMaOsxn8c
— Nathan Bennett (@vNerdyNate) August 26, 2019