Yunyuanyuan (10) | introduction to kubernetes in kubernetes

​Kubernetes brief introduction

One 、 background

1、 Change of deployment mode

• The era of traditional deployment ：
    Run the application on the physical server   Cannot define resource boundaries for application   Cause resource allocation problems

• Virtualization deployment era ：
    As a solution , Introducing virtualization   Virtualization technology allows you to work on a single physical server  CPU  Running multiple virtual machines on （VM）  Virtualization allows applications to be in  VM  Isolation between , And provide a certain degree of security   Information about an application   Can't be accessed by another application at will .  Virtualization technology can make better use of resources on physical servers   Because you can easily add or update applications  , So you can achieve better scalability , Reduce the cost of hardware and so on .  Every  VM  It's a complete computer , Run all components on top of virtualized hardware , Including its own operating system .

• Container deployment era ：
    The container is similar to  VM, But you can share the operating system between applications （OS）.  Containers are considered lightweight .  Container and  VM  similar , Has its own file system 、CPU、 Memory 、 Process space, etc .  Because they are separated from the infrastructure , So you can cross clouds and  OS  Release version for migration .

• agility ：
  Agile application creation and deployment ： And use  VM  Mirror Comparison , It improves the convenience and efficiency of container image creation .

• timeliness ：
  Continuous development 、 Integration and deployment ： By fast and simple rollback （ Because the image is immutable ）, Support reliable and frequent   Container image building and deployment .

• decoupled ：
  Focus on the separation of development and operation and maintenance ： In the build / Create an application container image when publishing , Instead of deploying .  To separate the application from the infrastructure .

• Observability ：
  Observability not only displays information and metrics at the operating system level , You can also display the health of the application and other indicators .

• Cross platform ：
  Cross development 、 Testing and production environment consistency ： Running on a laptop is the same as running in the cloud .

• portable ：
  Portability across cloud and operating system distributions ： Can be found in  Ubuntu、RHEL、CoreOS、 Local 、 Google Kubernetes Engine  And run anywhere else .

• Simplicity ：
  Application centric Management ： Raise the level of abstraction , From running on virtual hardware  OS  To use logical resources in  OS  Run the application on .

• Big distributed ：
  Loose coupling 、 Distributed 、 elastic 、 Liberation's microservice ： The application is broken down into smaller independent parts ,  And can be dynamically deployed and managed  -  Instead of running on a single large machine .

• Isolation, ：
  Resource isolation ： Predictable application performance .

• Efficiency ：
  Resource utilization ： High efficiency and
  high-density

2、 Containerization problem

• Flexible container application management

• Strong failover capability

• High performance load balancing access mechanism

• Convenient expansion

• Automated resource monitoring

3、 Why  Kubernetes

• Service discovery and load balancing
   Kubernetes  have access to  DNS  Name or own  IP  Address disclosure container , If the flow into the container is large , Kubernetes  Can load balance and distribute network traffic , So that the deployment is stable .

• Storage choreography
   Kubernetes  Allows you to automatically mount the storage system of your choice , For example, local storage 、 Public cloud providers, etc .

• Automatic deployment and rollback
    You can use  Kubernetes  Describes the required state of the deployed container , It can transfer the actual state at a controlled rate   Change to the desired state . for example , You can automate  Kubernetes  To create a new container for your deployment ,  Delete existing containers and use all their resources for new containers .

• Automatically complete the packing calculation
   Kubernetes  Allows you to specify what each container needs  CPU  And memory （RAM）.  When the container specifies a resource request ,Kubernetes  Better decisions can be made to manage the container's resources .

• Self repair
   Kubernetes  Restart the failed container 、 Replace the container 、 Kill does not respond to user-defined   Health check container , And don't advertise the service to the client until it's ready .

• Key and configuration management
   Kubernetes  Allows you to store and manage sensitive information , For example, password 、OAuth  Token and  ssh  secret key .  You can deploy and update the key and application configuration without rebuilding the container image , There is no need to expose the key in the stack configuration

4、 market share

Two 、 brief introduction

1、Kubernetes It's not something

• Kubernetes  Not traditional 、 All encompassing  PaaS（ Platform as a service ） System .

• Kubernetes  Run at the container level, not at the hardware level

• It provides  PaaS  Some common functions of the product ,  For example, deployment 、 Expand 、 Load balancing 、 Logging and monitoring .

• however ,Kubernetes  It's not a single system , The default solutions are all optional and pluggable . Kubernetes  Provides the foundation for building a developer platform , But it keeps the user's choice and flexibility where it matters .

• No restrictions on supported application types
  . Kubernetes  Designed to support extremely diverse workloads , Including statelessness 、 There are state and data processing workloads .  If the application can run in a container , Then it should be able to be in  Kubernetes  It works well on the Internet .

• Do not deploy source code
  , And don't build your app . 
  Continuous integration (CI)、 Delivery and deployment （CI/CD）
  Workflow depends on the culture and preferences of the organization as well as the technical requirements .

• No application level services as built-in Services
  , For example, middleware （ for example , Message middleware ）、  Data processing framework （ for example ,Spark）、 database （ for example ,mysql）、 cache 、 Cluster storage system  （ for example ,Ceph）. Such a component can be used in  Kubernetes  Up operation , also / Or it can be run by  Kubernetes  The application on the server is portable （ for example , 
  Open service agents
  ） To visit .

• No logging required 、 Monitoring or alerting solutions
  .  It provides some integration as a proof of concept , It also provides a mechanism for collecting and exporting indicators .

• No configuration language is provided or required / System
  （ for example  jsonnet）, It provides
  Declarative  API
  ,  The declarative  API  It can be any form of declarative specification .RESTful; Write yaml file

• Do not provide or use any comprehensive machine configuration 、 maintain 、 Manage or self repair systems .

• Besides ,Kubernetes  It's not just a choreography system , In fact, it eliminates the need for choreography .  The technical definition of choreography is to perform the defined workflow ： First, execute  A, And then execute  B, Re execution  C.  by comparison ,Kubernetes  Contains a set of independent 、 A composable control process ,  These processes continuously drive the current state to the desired state provided .  How to go from  A  To  C  It doesn't matter how you do it , There's no need for centralized control , This makes the system easier to use   And more powerful 、 More robust system 、 More resilient and scalable .