©2020-2024 ioDraw All rights reserved
large Outline
* Kubernetes Pod Detailed explanation
* Kubernetes Introduction to workload and services
Kubernetes Pod Detailed explanation
Kubernets overview
Kubernets overview
Kubernetes Key concepts -Pod
Kubernetes Key concepts -Pod
* stay Kubernetes in , pods Is able to create , dispatch , And management of the minimum deployment unit , Is a collection of containers , Instead of a separate application container
* The same Pod Containers in share the same network namespace , IP Address and port space .
* In terms of life cycle , Pod It's short-term, not long-term . Pods Is scheduled to the node , Remain on this node until it is destroyed .
Pod Detailed explanation - container
Pod Detailed explanation - container
Infrastructure Container： Basic container
* User not visible , No need to feel
* Maintain the whole Pod Cyberspace
InitContainers： Initialize container , It is generally used for service waiting and registration Pod Information, etc
* Execute before business container
* Sequential execution , Exit successfully （ exit 0）, Start the business container after successful execution
Containers： Business container
* Parallel start , After successful startup Running
Basic composition of container
Basic composition of container
Pod Detailed explanation - health examination
Pod Detailed explanation - health examination
Pod Detailed explanation - External input
Pod Detailed explanation - External input
configuration file （ ConfigMap） And key （ Secret） introduce
Pod Detailed explanation - Persistent storage
PV/PVC introduce
Pod Detailed explanation - Service domain name discovery
dnsPolicy： Pod The strategy of domain name resolution in
* ClusterFirst： use kube-dns As domain name resolution server
* Default： Using nodes （ kubelet） The specified domain name server resolves the domain name
* ClusterFirstWithHostNet： When Pod Use when using host network deployment
Kubernetes Introduction to workload and services
Pod Relationship to workload
Critical workload -ReplicaSet
* ReplicaSet For resolution pod The problem of capacity expansion and reduction .
* Usually used for stateless applications
Critical workload -Deployment
Critical workload -Deployment
Kubernetes Deployment Official updates are provided Pod and Replica
Set（ Next generation ReplicationController） Methods , You can use the Deployment Object only describes the desired state （ Expected operating state ）,Deployment The controller converts the current actual state to the desired state for you ;
Deployment Integrated online deployment , Rolling upgrade , Create a replica , Suspend online task , Resume online task , Roll back to a previous version （ success / stable ）Deployment Etc , To a certain extent ,
Deployment Can help us achieve unattended online , Greatly reduce the complexity of our online process communication , Operational risk .
Deployment Typical use cases for ：
* use Deployment To start （ go online / deploy ） One Pod perhaps ReplicaSet
* Check one Deployment Is the execution successful
* to update Deployment To recreate the corresponding Pods（ for example , A new one is needed Image）
* If existing Deployment instable , So roll back to an early stable Deployment edition
Critical workload -StatefulSet
Critical workload -StatefulSet
StatefulSet— Stateful application
It is used to solve various problems pod Instance independent lifecycle management , Provides the starting order and uniqueness of each instance
* stable , Unique network identifier .
* stable , Persistent storage --StatefulSet: each pod Corresponding to one pv
* Orderly , Elegant deployment and expansion .
* Order , Graceful deletion and termination .
* Ordered auto scrolling updates .
Critical workload -DaemonSet
Critical workload -DaemonSet
DaemonSet Can make all （ Or something specific ） Of Node Nodes run the same pod. When a node joins the kubernetes collection \ In a group , pod Will be （
DaemonSet） It is scheduled to run on this node , When node from kubernetes Removed from cluster , cover （
DaemonSet） Scheduled pod Will be removed , If deleted DaemonSet, All with this DaemonSet dependent pods Will be deleted .
in use kubernetes To run the application , A lot of times we need to be in one area （ zone） Or all Node Run the same daemons on （ pod）, For example, the following scenario ：
* each Node Runs a distributed storage daemons on , for example glusterd, ceph
* Run the log collector at each Node upper , for example fluentd, logstash
* The collection end of operation monitoring is in each Node, for example prometheus node exporter, collectd etc.
Critical workload -Job
Critical workload -Job
Kubernetes Key concepts -CustomResourceDefinition
Pod Relationship with services
Service
* Service Defined pods And the policy for accessing this set . Pods Set is defined by Service Provided by Label Selector completed
* Service The introduction of the pod Is transparent to the client , The client only needs to know service Address of , from service To provide an agent
* Service The abstraction makes the front-end client and the back-end Pods To decouple
* support ClusterIP, NodePort as well as LoadBalancer Three types
* Service The underlying implementation of userspace, iptables and ipvs Three modes
Service type ： ClusterIP, NodePort, LoadBalancer
Service type ： ClusterIP, NodePort, LoadBalancer
Ingress
* Ingress be based on service realization 7 Layer routing and forwarding capability
Technology