Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Maximize Architectural and Delivery Agility on Kubernetes

2,177 views

Published on

Rushi Ns, Senior Architect at SAP details best practices and guidelines for disaster recovery and high availability.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Maximize Architectural and Delivery Agility on Kubernetes

  1. 1. INTERNAL Dr. Rushi, NS SAP Nov 15 , 2018 Maximize Architectural and Delivery Agility Kubernetes Kubernetes
  2. 2. INTERNAL o JUST TO BE SURE à YOU LISTEN o I am not here for any product promotion(SAP or kubernetes Distritbutions) o All the information I provide here are the best of my knowledge based on the projects I have Architected and some information from Kubernetes source. Kubernetes
  3. 3. 3INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ 4 Pillars of K8S A foundation for right platform Best Practices High Availability Disaster Recovery AGENDA Kubernetes
  4. 4. 4INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ why Kubernetes? o Kubernetes automates quite a significant work in the cloud ->> o Kubernetes and Docker introduce a well-defined contract for all the cloud tenants o Kubernetes can be called an attempt to pretend that your application runs on one host via API Server. o Kubectl provides all means to starts/stop, diagnose and modify anything on your “virtual computer”, that is in fact just a container orchestrations. o Kubernetes unifies Infrastructure o Write Once, Run Anywhere o Workload Portability o Extend (own) Kubernetes, where sensible. E.g. to implement dynamic behavior o Life on an API calls (Pod, RC) o Kubernetes provides a container-centric management environment. o It orchestrates computing, networking, and storage infrastructure on behalf of user workloads. (It does not orchestrate workflows.) o This provides much of the simplicity of Platform as a Service (PaaS) with the flexibility of Infrastructure as a Service (IaaS), and enables portability across infrastructure providers. Kubernetes
  5. 5. 5INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ The right platform Kubernetes
  6. 6. 6INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ WorkloadsData Centers Infrastructures In-Memory Store Concept More Abstracted Blurring of IaaS, PaaS and SaaS Demarcations Less Abstracted Bare metal Virtual IaaS PaaS Serverless Certified hardware or White Boxes Vmware, XEN, LPAR AWS, Openstack, Azure, Google Cloud, kubernetes Cloud Foundry Openshift AWS Lambda, Azure, Google Functions, IBMwhisk,webtask, OracleFN,fission,kubeless, openfaas, openLambda Implementation Apps, DBS RunTime Apps, DBs RunTime Apps, DBs, Plugins RunTime Plugins, Code RunTime Code Events, Triggers Kubernetes exemplifies a unique architecture design of extensibility across all layers. kubernetes
  7. 7. 7INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Application Code Application Logs Library updates Overprovision ing Routing Container versions Scaling Rules CPU Utilization RAM Usage Node Placement Replication strategy Runtime upgrade OS Updates Hardware Network Switches, Racks Host migration FOCUS OF KUBERNETES Bare Metal VMs Containers Serverless GLDSkubernetes
  8. 8. 8INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Foundation of Right K8S Platform Workflow to self-heal K8S nodes (Approve/Reject) CI/CD integrated Storage Integrated Load balancers Integrated Security integrated (service Mesh) Network policies Data Governance k8S Distributions SUSE Caasp , Redhat Opensthift, Coreos, Pivotal, Rancher, MESOS, etc Infrastructure as a code MultiCloud Supported Support Baremetal Virtual Hypervisors Hyperconverged Fully Automated Deployment CIODemands Deploy & Operate Monitoring Tools GLDS Backup/Restore High Availability Disaster Recovery kubernetes Suse Caasp Offers workflow based cluster install
  9. 9. 9INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ SUSE CAASP kubernetes
  10. 10. 10INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ K8S Cluster Provisions Hyperconverged virtual Hypervisors baremetal Master/workers Network Storage Load balancer POD POD container container Master/Worker Network Storage Load balancer MulticloudFederation kubernetes
  11. 11. 11INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Best Practices kubernetes
  12. 12. 12INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Software Defined Approach Volumes (SDS) Overlay Network ETCD Artifactory (Registry) Meetup Specific Content EMC ScaleIO, VMWARE VSAN, Netapp ontap-Select, openEBS, Gluster, GPS/NFS 3 Member ETCD Cluster (out of Kubernetes control plane) Cluster HA Setup Flannel Calico
  13. 13. 13INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ High Availability No Single Point of Failure Meetup Specific Content Spread load across clusters , DNS servers and load balancers, federation minimizes the impact of cluster failure. Avoiding provider lock-in:, federation prevents cluster provider lock-in.
  14. 14. 14INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Dual Data Center Approach Multi Clusters Leader Elect Cloud Agnostic Meetup Specific Content 1. Deploy Master/Workers in secondary Data centers (on-premise to cloud vendors) 2. port-channel (LACP) 3. Data LIFS (Provision PV volumes from different SVM – SVM1 /SVM2 (NetApp or NFS ) 4. Load Balancers (F5/ Active-Active) 5. Artifcatory HA (Active/Active) 1. HAProxy provides access to the Kubernetes API 2. --leader-elect 3. master replicas in different zones 1. Multi Cloud (Gardner ) 2. Global Load Balancing (Ingress ) 3. GKE on-prem (in beta now)
  15. 15. 15INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ DC1 Storage LAN Storage LAN External Cluster LAN 1 External Cluster LAN 2 Deployment (PXE Boot) Server Kubernetes Master Node Kubernetes Minion Node LAN Deploy LAN Deploy LAN DC2 Clustered Router Clustered Loadbalancer K8S High Availability Network Topology Virtual LAN 1 Virtual LAN 2 Virtual LANs kubernetes
  16. 16. 16INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Networking Topology POD APP 1 Storage LAN 10.x.x.x/24 docker0 172.17.0.1/26 External LAN 10.x.x.x/24 NODE 1 flanneld etcd Deployment LAN 192.168.1.20/24 Internal LAN 192.168.2.20/24 cni0 172.30.0.1/24 flannel0 172.30.0.1/16 POD APP 2 Kube-proxy POD APP 1 Storage LAN 10.x.x.x/24 docker0 172.17.0.1/26 External LAN 10.x.x.x/24 NODE 2 flanneld etcd Deployment LAN 192.168.1.20/24 Internal LAN 192.168.2.20/24 cni0 172.30.1.1/24 flannel0 172.30.0.1/16 POD APP 2 Kube-proxy GLDSkubernetes
  17. 17. 17INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Disaster Recovery Cluster Federation Federated Ingress Cross Cluster communication kubernetes
  18. 18. 18INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ DR Solutions Federation ETCD Backup Resources Meetup Specific Content 1. Kubefed : federation control plane 2. Provisioning storage for etcd 3. CoreDNS 1. ETCD operator 2. ETCD Snapshot 1. ETCDCTL_API=3 etcdctl --endpoints $ENDPOINT snapshot save snapshot.db 3. ETCD Member Cluster 1. ETCDCTL_API=3 etcdctl snapshot restore snapshot.db -- name m1 --initial-cluster m1=http://host1:2380,m2=http://host2:2380,m3=http://host3:238 0 --initial-cluster-token etcd-cluster-1 --initial-advertise-peer- urls http://host1:2380 1. cluster resources and persistent volumes.
  19. 19. 19INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Final Words § Compare the models usability and roadmap with K8S § Team awareness and upskills on latest & greatest. § “infrastructure as a code” with fully automated CI/CD pipelines for cluster install/upgrade. § Client-go for operators development § Network and Storage DR should be ready before cluster install § HA & DR in place after cluster install § Private Registry HA (JFROG Artifacatory) Meetup Specific Content
  20. 20. 20INTERNAL© 2017 SAP SE or an SAP affiliate company. All rights reserved. ǀ Reference Materials § JFROG Artificatory HA https://www.jfrog.com/confluence/display/RTF/HA+Installation+and+Setup § Think like kubermesh solution https://github.com/kubermesh/kubermesh § Kubernetes as a service: https://github.com/gardener/gardener § Reference: HA Compatible Cluster – https://kubernetes.io/docs/tasks/administer-cluster/highly-available-master/ § Kubernetes Solutions: https://kubernetes.io/docs/setup/pick-right-solution/ § Backup Cluster Resources and Persistent Volumes https://github.com/heptio/ark § Cluster Federation and Global Load Balancing – https://medium.com/google-cloud/planet-scale-microservices-with-cluster-federation-and-global-load- balancing-on-kubernetes-and-a8e7ef5efa5e § Master Replica info https://kubernetes.io/docs/tasks/administer-cluster/highly-available-master/#best-practices- for-replicating-masters-for-ha-clusters Automate HA deployments ▫ https://github.com/kubernetes/community/blob/master/contributors/design-proposals/cluster-lifecycle/ha_master.md § ETCD operator https://github.com/coreos/etcd-operator § Hyperconverged Kubernetes “Robin “https://robin.io/get-robin/ Meetup Specific Content
  21. 21. Thank you. Contact information: Dr. Rushi NS (rushi.ns@sap.com) SAP , Palo Alto, CA kubernetes

×