Shulou(Shulou.com)05/31 Report--

This article mainly introduces the relevant knowledge of "openstack cloud computing cinder architecture and what are the functions of each component". The editor shows you the operation process through an actual case. The method of operation is simple and fast, and it is practical. I hope that this article "openstack cloud computing cinder architecture and what are the functions of various components" can help you solve the problem.

1. Cinder components 1. Cinder is mainly composed of # cinder-api#cinder-scheduler#cinder-volume2, cinder components function:

Cinder-api is the endpoint of the cinder service, which provides the rest interface, handles client requests, and sends RPC requests to the cinder-scheduler component.

Cinder-scheduler is responsible for scheduling cinder requests, and its core part is scheduler_driver. As the driver of scheduler manager, it is responsible for the specific scheduling processing of cinder-volume, sending cinder RPC requests to the selected cinder-volume.

Cinder-volume is responsible for the specific processing of volume requests, and different backend stores provide volume storage space. At present, major storage manufacturers have actively contributed the driver of storage products to the cinder community.

3. The working principles of nova and cinder are similar

The main components of nova are:

# nova-api#nova-scheduler#nova-compute#nova-conductor 2 cinder Architecture Diagram

Communication between openstack components: call the rest interface provided by each component api. Intra-component communication is based on rpc (remote procedure call) mechanism, while rpc mechanism is implemented based on AMQP model.

From the point of view of the use of rpc, the processes of nova,neutron and cinder are similar. We take cinder as an example to illustrate the rpc mechanism.

Reference link: https://developer.ibm.com/

Three RPC mechanisms

The implementation of the RPC (Remote Producer Call) mechanism within the Openstack component is based on AMQP (Advanced Message Queuing Protocol) as the communication model, so as to meet the loose coupling within the component. AMQP is a message middleware protocol for asynchronous message communication.

The AMQP model has four important roles:

Exchange: forward messages to the corresponding Message Queue according to Routing key

Routing key: used by Exchange to determine which messages need to be sent to the corresponding Message Queue

Publisher: the sender of the message that sends the message to the Exchange and indicates the Routing Key so that the Message Queue can receive the message correctly

Consumer: message recipient, getting message from Message Queue

The message publisher Publisher sends Message to Exchange and explains Routing Key. Exchange is responsible for routing according to the Routing Key of the Message and correctly forwarding the Message to the appropriate Message Queue. Consumer listening on Message Queue will read messages from Queue.

Routing Key is the basis for Exchange to forward information, so each message has a Routing Key indicating that the destination address of the message can be accepted, and each Message Queue can binding by telling Exchange the Routing Key it wants to receive, so that Exchange can correctly forward the message to the corresponding Message Queue.

Publisher can be divided into four categories:

Direct Publisher sends peer-to-peer messages

Topic Publisher uses publish-subscribe mode to send messages

The transmission of Fanout Publisher broadcast messages

Notify Publisher and Topic Publisher send messages related to Notification.

Exchange can be divided into three categories:

1.Direct Exchange matches exactly according to Routing Key, and only the corresponding Message Queue will receive the message.

2.Topic Exchange performs pattern matching according to Routing Key, and any Message Queue that matches the pattern match will receive a message.

3.Fanout Exchange forwards the message to all bound Message Queue.

AMQP message model

RPC sends request

The Client side sends the RPC request. The publisher sends the message and declares the message address. The consumer receives the message and processes the message. If the message reply is needed, it returns the result message of the processing request.

The OpenStack RPC module provides three RPC calling methods, rpc.call,rpc.cast and rpc.fanout_cast, to send and receive RPC requests.

1.rpc.call sends the RPC request and returns the request processing result. The request processing flow is shown in figure 5. The Topic Publisher sends the message, and the Topic Exchange forwards the message to the corresponding Message Queue according to the message address. The Topic Consumer listens to the Message Queue and processes the message when it finds the message to be processed. The Direct Publisher processes the request result message and requests the sender to create the return result of the Direct Consumer listening message.

Figure 5.RPC.call message flow

The RPC request sent by 2.rpc.cast has no return, and the request processing flow is shown in figure 6. Unlike rpc.call, there is no need for the return of the request processing result, so there is no Direct Publisher and Direct Consumer processing.

Figure 6. RPC.cast message flow

3.rpc.fanout_cast is used to send RPC broadcast messages and no result is returned.

Figure 7. RPC.fanout message processing

This is the end of the introduction to "what is the architecture of openstack cloud computing cinder and the functions of its components". Thank you for reading. If you want to know more about the industry, you can follow the industry information channel. The editor will update different knowledge points for you every day.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.