KEMBAR78
Deploying kubernetes at scale on OpenStack | PDF
DEPLOYING KUBERNETES AT
SCALE ON OPENSTACK
Victor Palma – OpenStack Architect
TONIGHT’S AGENDA
What to consider when
deploying Kubernetes?
OpenStack Components
Network Considerations
WHAT IS KUBERNETES?
KUBERNETES IS __
• Kubernetes is an open-source “platform”
for automating deployment, scaling, and
operations of application containers
across clusters of hosts, providing
container-centric infrastructure.
• portable: public, private, hybrid, multi-
cloud
• extensible: modular, pluggable, hook
able, compassable
• self-healing: auto-placement, auto-
restart, auto- replication, auto-scaling
KUBERNETES
HISTORY
• Borg, Omega and the
Origin of Kubernetes
• The name Kubernetes
originates from Greek,
meaning "helmsman" or
"pilot", and is the root of
"governor" and
"cybernetic".
• K8s is an abbreviation
derived by replacing the 8
letters "ubernete" with 8.
MASTER COMPONENTS
6
API
UI
CLI
etcd
API Server Scheduler Controller
NODE COMPONENTS
7
Kubernetes
Master
Docker kubelet
Daemon sets
Kube-proxy
fluentd
Pod
Pod
PodAddons
Pod Pod
Pod
Pod
Pod
Pod
UIDNS
Node
WHAT KIND OF
EXPERIENCE DO YOU
WANT TO PROVIDE?
WHO?
• Developers
• Operations
WHAT?
• Self Healing
• Auto Scaling
• Rollout control
• Resource
Management
• Service Discovery
• Service Mesh
• Federated
WHERE?
• Public Clouds
• Private Cloud
WHAT TO CONSIDER
WHEN DEPLOYING
KUBERNETES
STORAGE, NETWORKING, SERVICES
13
STORAGE
• Local Storage
vs
• Shared Storage
K8’S
NETWORKING
• Highly-coupled
container-to-container
communications
• Pod-to-Pod
communications
• Pod-to-Service
communications
• External-to-Service
communications
KUBERNETES NETWORKING ADDRESSES
16
• Cross-node pod-to-pod connectivity (east-west traffic)
• Services running in the pods are discovered by other
pods
• Pod-to-Pod traffic is load balanced when consuming a
service.
• Expose Services for external Consumption (for north-
south traffic).
• Network Segmentation
• Global load balancing via multi-cluster federation
KUBERNETES CNI PLUGINS
17
• ACI
• Big Cloud Fabric
• Calico
• Cilium
• Contiv
• Contrail
• Flannel
• Kube-router
• L2 + Linux Bridging
• OpenVswitch
• OVN
• Weave net
LAYER 2
(SWITCHING)
• Simple
• Bridge CNI plug-in
LAYER 3
(ROUTING)
• Scalable
• CNI plug-in: BGP
‣ Calico
‣ Contiv
‣ etc
OVERLAY
• Simple
• CNI plug-in: (vxlan, encapusulation)
‣ Weave
‣ flannel
SERVICE MESH
DEPLOYMENT
• Kubespray
• Kops
• Techtonic
• etc.
TOP THINGS TO CONSIDER
23
ETCD
• etcd
‣ Self Hosted
‣ Stand Alone Cluster
UPGRADE
STRATEGY
• OS Upgrades
• Kubernetes
Upgrades
SERVICES YOU
PROVIDE
• Monitoring
‣ Prometheus
• Alerting
‣ Grafana
• Logging
• RBAC
LET’S PUT
EVERYTHING
TOGETHER
SOMETIMES YOU CAN’T START FROM
SCRATCH!
29
NOT BORN IN THE CLOUD!
• UI Layer
• Business Logic Tier
Data
Access Tier
Integration
Tier
MainframeRDMS
DB(s)
• Monolithic Apps
• Dependent on Heavy, stagnant tech
• Configuration is embedded and static
• Requires specific infrastructure and configuration
• Inability to decompose deployments
• Lack of portability
31
OPENSTACK PROJECTS
• Core Services:
▸ NOVA – Compute
▸ NEUTRON – Networking
▸ SWIFT – Object Storage
▸ CINDER – Block Storage
▸ KEYSTONE – Identity
▸ GLANCE – Imaging Service
•Optional Services:
▸Horizon – Dashboard
▸Octavia – Load balancing
▸Ceilometer – Telemetry
▸Heat – Orchestration
▸Trove – Database
▸Sahara – Elastic Map Reduce
▸Ironic – Bare-Metal Provisioning
▸Zaqar – Messaging Service
▸Manila – Shared FS
▸Designate – DNS Service
▸Barbican – Key Management
▸Magnum – Containers
▸Murano – App Catalog
â–¸Congress - Governance
Core and Optional Services
32
THE KUBERNETES CLUSTER AND THE
SERVICES RUNNING INSIDE
Hardware
Image
Registry
Logging Monitoring Alerting RBAC
Applications
Operating Kubernetes
Traditional RPC support on hardware,
OpenStack, and Ceph.
Deploy, configure and manage Kubernetes on
OpenStack in a highly available way. We've done
the work to integrate Kubernetes with
OpenStack.
Deploy, configure and manage per cluster
services for your application.
Kubernetes is a new technology so that means
that you have to manage cluster operations,
deployments, and Kubernetes best practices.
Hardware
OpenStack New Alpha Features
Changes in 1.9
KUBERNETES
33
• Block Storage (Cinder) V3 is now
supported.
• Load Balancer (Octavia) V2 is now
supported, in addition to Neutron
LBaaS V2.
• Additional security group updates
‣ Kubernetes will now automatically
determine the security group for the node
‣ Nodes can now belong to multiple security
groups
• Neutron LBaas V1 support has been
removed.
• Support for IPv6
• Alpha support for CoreDNS as
a drop-in replacement for kube-
dns
• networking.k8s.io/v1 to replace
extensions/v1beta1
NetworkPolicy API
• Container Storage Interface
(CSI)
• Raw block volumes
RACKSPACE OFFERS MANAGED
KUBERNETES FOR ENTERPRISE IT
• Optimized for cost efficiency
• Best-in-class compute
• Kubernetes expertise from certified cloud
specialists
• Centralized logging and monitoring
• Upgrades and patching
• Isolated multi-tenant or single tenant
• Security-hardened environments
• Full turn-key solution with 1-click deployment
• End-to-end tutorials and training
• Native access to Kubernetes/Docker ecosystems
• Unified GUI and tooling
• Delivered via OpenStack
• Full OpenStack integration
• User management, identity, quotas and limits
Operating Expertise Ease of Use
Management Technology Integration
Rackspace
Managed
Kubernetes
34
Fully Managed Kubernetes: Secure and highly performant with an intuitive user experience
BETA AVAILABLE NOW!
Copyright © 2016 Rackspace | Rackspace® Fanatical Support® and other Rackspace marks are either registered service marks or service marks of Rackspace US, Inc. in the United States and other countries. Features, benefits and pricing
presented depend on system configuration and are subject to change without notice. Rackspace disclaims any representation, warranty or other legal commitment regarding its services except for those expressly stated
in a Rackspace services agreement. All other trademarks, service marks, images, products and brands remain the sole property of their respective holders and do not imply endorsement or sponsorship.
THANK YOU
ONE FANATICAL PLACE | SAN ANTONIO, TX 78218
US SALES: 1-800-961-2888 | US SUPPORT: 1-800-961-4454 | WWW.RACKSPACE.COM

Deploying kubernetes at scale on OpenStack

  • 1.
    DEPLOYING KUBERNETES AT SCALEON OPENSTACK Victor Palma – OpenStack Architect
  • 2.
    TONIGHT’S AGENDA What toconsider when deploying Kubernetes? OpenStack Components Network Considerations
  • 3.
  • 4.
    KUBERNETES IS __ •Kubernetes is an open-source “platform” for automating deployment, scaling, and operations of application containers across clusters of hosts, providing container-centric infrastructure. • portable: public, private, hybrid, multi- cloud • extensible: modular, pluggable, hook able, compassable • self-healing: auto-placement, auto- restart, auto- replication, auto-scaling
  • 5.
    KUBERNETES HISTORY • Borg, Omegaand the Origin of Kubernetes • The name Kubernetes originates from Greek, meaning "helmsman" or "pilot", and is the root of "governor" and "cybernetic". • K8s is an abbreviation derived by replacing the 8 letters "ubernete" with 8.
  • 6.
  • 7.
    NODE COMPONENTS 7 Kubernetes Master Docker kubelet Daemonsets Kube-proxy fluentd Pod Pod PodAddons Pod Pod Pod Pod Pod Pod UIDNS Node
  • 8.
    WHAT KIND OF EXPERIENCEDO YOU WANT TO PROVIDE?
  • 9.
  • 10.
    WHAT? • Self Healing •Auto Scaling • Rollout control • Resource Management • Service Discovery • Service Mesh • Federated
  • 11.
  • 12.
    WHAT TO CONSIDER WHENDEPLOYING KUBERNETES
  • 13.
  • 14.
  • 15.
  • 16.
    KUBERNETES NETWORKING ADDRESSES 16 •Cross-node pod-to-pod connectivity (east-west traffic) • Services running in the pods are discovered by other pods • Pod-to-Pod traffic is load balanced when consuming a service. • Expose Services for external Consumption (for north- south traffic). • Network Segmentation • Global load balancing via multi-cluster federation
  • 17.
    KUBERNETES CNI PLUGINS 17 •ACI • Big Cloud Fabric • Calico • Cilium • Contiv • Contrail • Flannel • Kube-router • L2 + Linux Bridging • OpenVswitch • OVN • Weave net
  • 18.
  • 19.
    LAYER 3 (ROUTING) • Scalable •CNI plug-in: BGP ‣ Calico ‣ Contiv ‣ etc
  • 20.
    OVERLAY • Simple • CNIplug-in: (vxlan, encapusulation) ‣ Weave ‣ flannel
  • 21.
  • 22.
  • 23.
    TOP THINGS TOCONSIDER 23
  • 24.
    ETCD • etcd ‣ SelfHosted ‣ Stand Alone Cluster
  • 25.
  • 26.
    SERVICES YOU PROVIDE • Monitoring ‣Prometheus • Alerting ‣ Grafana • Logging • RBAC
  • 27.
  • 29.
    SOMETIMES YOU CAN’TSTART FROM SCRATCH! 29
  • 30.
    NOT BORN INTHE CLOUD! • UI Layer • Business Logic Tier Data Access Tier Integration Tier MainframeRDMS DB(s) • Monolithic Apps • Dependent on Heavy, stagnant tech • Configuration is embedded and static • Requires specific infrastructure and configuration • Inability to decompose deployments • Lack of portability
  • 31.
    31 OPENSTACK PROJECTS • CoreServices: ▸ NOVA – Compute ▸ NEUTRON – Networking ▸ SWIFT – Object Storage ▸ CINDER – Block Storage ▸ KEYSTONE – Identity ▸ GLANCE – Imaging Service •Optional Services: ▸Horizon – Dashboard ▸Octavia – Load balancing ▸Ceilometer – Telemetry ▸Heat – Orchestration ▸Trove – Database ▸Sahara – Elastic Map Reduce ▸Ironic – Bare-Metal Provisioning ▸Zaqar – Messaging Service ▸Manila – Shared FS ▸Designate – DNS Service ▸Barbican – Key Management ▸Magnum – Containers ▸Murano – App Catalog ▸Congress - Governance Core and Optional Services
  • 32.
    32 THE KUBERNETES CLUSTERAND THE SERVICES RUNNING INSIDE Hardware Image Registry Logging Monitoring Alerting RBAC Applications Operating Kubernetes Traditional RPC support on hardware, OpenStack, and Ceph. Deploy, configure and manage Kubernetes on OpenStack in a highly available way. We've done the work to integrate Kubernetes with OpenStack. Deploy, configure and manage per cluster services for your application. Kubernetes is a new technology so that means that you have to manage cluster operations, deployments, and Kubernetes best practices. Hardware
  • 33.
    OpenStack New AlphaFeatures Changes in 1.9 KUBERNETES 33 • Block Storage (Cinder) V3 is now supported. • Load Balancer (Octavia) V2 is now supported, in addition to Neutron LBaaS V2. • Additional security group updates ‣ Kubernetes will now automatically determine the security group for the node ‣ Nodes can now belong to multiple security groups • Neutron LBaas V1 support has been removed. • Support for IPv6 • Alpha support for CoreDNS as a drop-in replacement for kube- dns • networking.k8s.io/v1 to replace extensions/v1beta1 NetworkPolicy API • Container Storage Interface (CSI) • Raw block volumes
  • 34.
    RACKSPACE OFFERS MANAGED KUBERNETESFOR ENTERPRISE IT • Optimized for cost efficiency • Best-in-class compute • Kubernetes expertise from certified cloud specialists • Centralized logging and monitoring • Upgrades and patching • Isolated multi-tenant or single tenant • Security-hardened environments • Full turn-key solution with 1-click deployment • End-to-end tutorials and training • Native access to Kubernetes/Docker ecosystems • Unified GUI and tooling • Delivered via OpenStack • Full OpenStack integration • User management, identity, quotas and limits Operating Expertise Ease of Use Management Technology Integration Rackspace Managed Kubernetes 34 Fully Managed Kubernetes: Secure and highly performant with an intuitive user experience BETA AVAILABLE NOW!
  • 36.
    Copyright © 2016Rackspace | Rackspace® Fanatical Support® and other Rackspace marks are either registered service marks or service marks of Rackspace US, Inc. in the United States and other countries. Features, benefits and pricing presented depend on system configuration and are subject to change without notice. Rackspace disclaims any representation, warranty or other legal commitment regarding its services except for those expressly stated in a Rackspace services agreement. All other trademarks, service marks, images, products and brands remain the sole property of their respective holders and do not imply endorsement or sponsorship. THANK YOU ONE FANATICAL PLACE | SAN ANTONIO, TX 78218 US SALES: 1-800-961-2888 | US SUPPORT: 1-800-961-4454 | WWW.RACKSPACE.COM