Why does the docker-containerd-shim process need to be started or run?
It allows the container to exit after the runtime starts. It is not necessary to run each container with container operations-even in the case of containerd and all the tricks of dockerd. The container’s standard IO and other file descriptors are also available-report the container’s exit status to container d.
What is the difference between rkt and containerd? One major difference is that RKT can be used in a production environment as a non-daemonic tool. Integration and implementation have key uses, such as special containers, the container CoreOS-Linux’s RKT uses the container to mirror the agent executed by Kubernetes in this way. With this kublet, more examples can be included in the Kubernetes ecological and economic environment. Using rkt to mount a volume as a social pressure containerization method, which means that rkt can be integrated and used with Linux’s init system analysis, because rkt’s own economic development problems are not the only This is for an init system. Kubnets supports the deployment of containers. The containers it supports are not limited to the dock, the core OSrkt, but also a container player. Although compared with dockers, the absolute style is still obvious, but it is always better than no competition.
The container is very light, even quantized technology, compared to the other in terms of physical and virtual machines. This means that more container instances can be created on the basis of the same resources. Once faced with large-scale applications distributed on multiple hosts with hundreds of containers, traditional or independent container management solutions will become unbearable Burdened. On the other hand, due to improved local support for microservices, container granularity is getting smaller and smaller in container clusters. In this case, it is necessary to manage the service in the container or micro, accept and enter the external environment in an orderly manner. In order to achieve scheduling, load balancing and task distribution, simple and effective management of container instances is rapidly increasing. Naturally developed into the main task of the container orchestration system.
The container cluster management tool manages a multi-container combination of applications on a set of servers, and each application cluster looks like a deployment or management entity of a container orchestration tool. The container cluster enterprise management research tool realizes full automation for the application cluster, including the deployment of technical application examples, application updates, health and safety checks, and flexible expansion. Automatic fault tolerance and so on…
Container orchestration and management system-Kubernetes Google’s open source container management system, derived from the long-existing internal Borg system, is used by many companies due to its rich functions. Its development route focuses on standardization and manufacturer “neutrality”. Supporting different container runtimes and bottom engines, gradually relying on Docker, and the core of Kubernetes is how to solve automatic deployment, expansion and management of containerized applications. The currently starred project on GitHub is 43K. Docker Swarm: After Docker 1.2, Swarm is integrated into the Docker engine. Users can quickly and easily create a docking container cluster. Almost fully compatible — docker — the project’s API — is characterized by the current number of stars on github at 5.3K. Mesosphere — — Marathon: The design goal of the scheduling research framework of Apache — — Mesos is to become a data information center operation system, which can completely take over the management accounting work of the data technology center. Completely take over the management and accounting work of the data information center, Mesos concept is the data technology center operation control system. In order to solve the network, computing and storage problems of the IaaS layer, the core of Mesos is to solve the problem of the physical resource layer. Marathon is a container orchestration platform designed for the middle layer — DC/OS and the Apache middle layer — DC/OSMesos.