Container Conf 2017: Rancher Kubernetes

Editor's Notes

• #5: The momentum of container adoption is undeniable: The average company QUINTUPLES its Docker usage within 9 MONTHS There are 460K Dockerized apps, a 3100% GROWTH over 2 years Docker containers have been downloaded more than 4 BILLION times
• #6: But running containers in production still isn’t easy… Growing number of tools with a high rate of change is significantly increasingly the complexity of building, deploying and updating containerized environments Increased density combined with decreased lifespans (containers lifespan 1/6 of a VM) significantly increases the volatility of containerized environment compared to a traditional VM environment - there is an order-of-magnitude increase in the number of things that need to be individually managed and monitored, which significantly increases the volatility of the environment.
• #7: Rancher is a complete container management platform that makes it easy to… Run containers in production with the most complete set of container and infrastructure management capabilities Manage applications by simplifying day to day application lifecycle management Innovate with containers without compromising flexibility by empowering developers with fast access to the latest tools And Rancher is production ready Enterprise-class features such as role-based access control, integration with LDAP and Active Directories, detailed audit logs, high-availability management servers and encrypted networking are available out of the box. With over 2.4 million downloads and optional 24x7x365 support, Rancher has quickly become the platform of choice for DevOps and IT teams who are serious about running containers in production More than 30 customers are using Rancher in production including large enterprise such as US Bank, IBM, Cisco, Invisalign and more.
• #9: There’s a slew of technologies that are in the ecosystem and they all serve useful purposes but it’s paralyzing having to deal w/all of the individually. Very flexible working w/all of these open source projects but it’s challenging updating and keeping in sync w/all of them. Rancher provides a turnkey container service and looked at all of the the technologies and taking ownership of them as it relates to being deployed at customer site. Rancher ties all of this together and makes it Easy to deploy, easy to support, easy to scale. Challenge is integrating it into your org and make use of the investments you’ve made
• #15: Environment templates are ways to create blueprints for different environments you need. You can customize the blueprint for various aspects such as storage, networking, DNS etc. In some environments you might want a fully HA setup where as in another environment you might be ok with a non HA setup. The way to achieve this is to create separate environment templates for different needs and then launch environments from templates. Of course you can launch more than one environment from same template. Some of stacks available in environment customization might be community supported and some are officially supported by Rancher - be sure to choose the right one. Now let’s quickly create an environments which we will use for the demo today (We will create another one in a bit). In the environment - we will configure the digitalocean-dns, healthcheck, kubernetes, network-services, portainer, ipsec, scheduler. Also we choose not to use resiliency planes in Kubernetes for simplicity, but we will discuss about it.
• #16: This slide will eventually be removed in the actual version of presentationa and a live demo will be given instead. This slide only demonstrates the rough flow to be demoed.
• #17: Resiliency Plaines allow you to build system that are highly available and fault tolerant. This is specially important when you have to build systems for production grade workloads. In Rancher, for a given environment template you can enable or disable resiliency planes. If you enable them you have to use nodes labels to identify the hosts belonging to certain plane. At a minimum it is good idea to have 3 node for etcd so that the data plane can tolerate failure of one node. If you need higher fault tolerance then you can go for 5 etcd nodes - which provides a tolerance of upto 2 hosts failing. For the Kubernetes or orchestrate layer, you need at least two nodes. On compute side you can have minimum 1 but in real world you might have many more hosts in compute plane.
• #18: Cloud provider configuration enables you with native integration with the cloud provider so you can seamlessly use some of underlying resources. At the moment rancher supports two cloud providers - one is rancher and other is AWS. There might be more in near future. With Rancher cloud provider you can get load balancing even without a cloud provider. The rancher provided load balacing is based on haProxy. With AWS you get native integration with EBS, ELB & EC2 instances. For a service type Load Balancer, an ELB is provsioned, although you can create a Rancher load balancer using Ingress.
• #19: We are using digital ocean demo for this one but you could also use the stack for Route53. The basic idea is that foe every service you want to expose outside of cluster, you simply create LoadBalncer type service and the DNS records are managed automatically based one semantics. And the format that the service name should put in DNS is customizable. So here is how it works - you of course need to have a domain name. Then you need to create a hosted zone(AWS)/cloud DNS (Google Cloud)/DOMAIN (Digital Ocean) for that domain name. Ensure that you update the name servers in your domain name registrar to point to correct name servers of cloud provider. Once this is done, then in the template we simply provide DO access key and the name of domain name that needs to be managed. Now as you create services - we will see two things: How when you create a service - the Load Balancer using HA Proxy is created (Which is what we discussed in previous slide) How a DNS entry is made in Digital Ocean Domain and updated with IP.
• #20: Rancher provides
• #24: In second environment, create a AWS cloud provider based Kubernetes environment. Then create hosts manually and apply them Instance Roles - separate once for master & agent machines. Then custom add them to Rancher.
• #25: This slide is temporary - and actual demo will be shown here.