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

This article mainly explains "what are the general designs and improvements in database resource delivery". The explanation in this article is simple and clear, and is easy to learn and understand. let's go deep into the editor's train of thought. Let's study and learn about the general design and improvement of database resource delivery.

For installation and deployment, the process involved is more complicated, and with the subsequent maintenance and management, the process will change. In the previous code layer maintenance, it will easily lead to difficult adaptation and process instability, resulting in a large gap between the delivery efficiency of installation and deployment and expectations.

The existing processes are as follows:

If the above process has the following problems, I believe that in many small and medium-sized companies will be more or less involved.

Overall performance

1) in the code implementation, the process is relatively bloated, hard-coded, and the risk of process change is high.

2) too much information in the resource application, the information is not concise enough, and it is not friendly to the business side.

3) the current resource flow is relatively complex and belongs to customized development. If there is a similar configuration in other processes, the code implementation reuse is low.

Resource approval

4) the delivery time of resources is longer than expected, which is reflected in the examination and approval link on the one hand and the high trial and error cost of resource delivery on the other hand.

5) the database resource application for the test environment is not supported in the work order at present, and the process of creating the database needs to be guided manually.

Host resource pool filtering

6) in resource delivery, if there is a resource configuration that does not match in the work order, it cannot be delivered, and the work order document needs to be modified.

7) at present, the link of the host resource pool is artificially controlled, which requires manual input of host information, and there is no threshold management and resource pre-application process of the resource pool.

Database resource delivery

8) if the process execution fails, the retry process detection is relatively weak, and some additional processing work needs to be done manually.

9) if the process is too long, the probability of error in a certain link is high, which leads to a high probability of error in the whole deployment.

10) with the access of the new version of the database, the original model is difficult to be compatible. At present, most of the new environments are deployed in manual mode.

11) if you apply for a single instance with one master and two slaves, and in a cluster environment, it cannot be supported and adapted.

Database permission delivery

12) the delivery of permissions after the delivery of resources may not be clear at the time of business resource application, so the probability of later change is high, but if you apply manually, you need to submit an automated online cooperation order (database building). Permission application cooperation order (need another round of approval), build table (automatic launch cooperation order or object operation cooperation order), developers who are not familiar with the process The process can be complex and unclear.

The corresponding improvement strategies and directions are as follows. In short, we hope that the basic work that accounts for more than 90% of the pre-application and pre-configuration of resources can be done first. When the business submits an application, DBA only needs to deal with that 10% of the configuration management.

Overall performance

1) in the code implementation, the process is relatively bloated, hard-coded, and the risk of process change is high.

Improvement strategy: process scheduling implementation based on configuration, considering process changes at the early stage of design, through multi-process configuration and choreography to support different business scenarios, such as support for single instance, one master and one slave, one master and two slaves, the process is similar but different, and multi-class requirements are realized by configuring different processes.

2) too much information in the resource application, the information is not concise enough, and it is not friendly to the business side.

Improvement strategy: optimize the current front-end configuration, remove unnecessary information and required entries, and reduce the number of entries by at least 20%.

3) the current resource flow is relatively complex and belongs to customized development. If there is a similar configuration in other processes, the code implementation reuse is low.

Improvement strategy: for process scheduling and task configuration, it can be realized through general configuration and general services to improve code reuse and stability construction.

Resource approval

4) the delivery time of resources is longer than expected, which is reflected in the examination and approval link on the one hand and the high trial and error cost of resource delivery on the other hand.

Improvement strategy:

For the resource delivery of the test environment, it is actually the database delivery, which can simplify the process implementation.

For the resource delivery of the development environment, the approval link can be removed directly and can be realized later by virtualizing the multi-tenant model.

At present, the existing examination and approval links are still retained for the resource delivery of online environment, and the embodiment of resource cost is open to question.

5) the database resource application of the test environment is not supported in the work order, and the process of creating the database needs to be guided manually.

Improvement strategy: as above

Host resource pool filtering

6) in resource delivery, if there is a mismatched resource configuration in the work order, it cannot be delivered, and the data of the work order needs to be modified.

Improvement strategy: the configuration of resource pools can be differentiated, but suitability needs to be considered. Resource configuration is implemented according to the criteria of priority scalability, such as database resources for business application 8C8G. Currently, there are five instance resources in the resource pool:

① 2 4C4G, 2 8C8G focus 1 8C16G, can be delivered according to the specification of 2 8C8G

If ② has 2 4C4G and 1 8C8G focus and 1 8C16G, it can be delivered according to the specification of 1 8C8G focus and 1 8C16G, in which 8C16G binds the main library first.

③ 2 4C4G, 1 8C8G focus 2 8C16G, can be delivered according to the specification of 2 8C16G

7) at present, the link of the host resource pool is artificially controlled, which requires manual input of host information, and there is no threshold management and resource pre-application process of the resource pool.

Improvement strategy: at the level of rapid delivery of resources, the resource layer can be divided into host resource pool and database instance resource pool, and built hierarchically through host resource pool and instance resource pool, in which the instance resource pool only retains the available resources. after the resources are used, they need to be archived to the resource history details, and the host resource pool needs to be docked with the system department through the process. For this host resource pool, we need to consider the implementation of threshold alarm and provide the necessary interfaces for the system department to call back.

Database resource delivery

8) if the process execution fails, the retry process detection is relatively weak, and some additional processing work needs to be done manually.

9) if the process is too long, the probability of error in a certain link is high, which leads to a high probability of error in the whole deployment.

10) with the access of the new version of the database, the original model is difficult to be compatible. At present, most of the new environments are deployed in manual mode.

11) if you apply for a single instance with one master and two slaves, and in a cluster environment, it cannot be supported and adapted.

Improvement strategy: at present, task details are configured through a general process. For task objects, the global uniqueness of pipelined numbering needs to be considered.

Database permission delivery

12) the delivery of permissions after the delivery of resources may not be clear at the time of business resource application, so the probability of later change is high, but if you apply manually, you need to submit an automated online cooperation order (database building). Permission application cooperation order (need another round of approval), build table (automatic launch cooperation order or object operation cooperation order), developers who are not familiar with the process The process can be complex and unclear.

Improvement strategy: for the processing of resource application documents, we can provide a more flexible support mode and reduce the submission of multiple work orders as much as possible.

The overall design of the general task flow is mainly layered in the following way.

Picture

It will cover less in more detail, such as task dependence, timeout processing, etc., mainly based on the basic process execution mode.

Among them, the orchestration layer realizes the arrangement of the process and the configuration of the process task, which involves the basic information and does not involve the specific implementation details.

The application layer is a business-independent data model, which needs to define a globally unique batch number (batch_no) in the business layer, which can also be understood as a globally unique object ID.

The task execution layer is mainly the implementation of general tasks, in which the configuration details of process tasks are based on the combination of application layer data configuration and process task configuration to form the registration of task details, such as when submitting deployment requests. It's the execution plan of the task details.

The process task detail log maintains the execution log and status of the process task details. If the task executes successfully, the corresponding task detail record status will be updated, otherwise, if it fails, you need to start the retry mechanism.

Thank you for your reading. the above is the content of "what are the general designs and improvements in database resource delivery". After the study of this article, I believe you have a deeper understanding of the general design and improvement of database resource delivery, and the specific use needs to be verified in practice. Here is, the editor will push for you more related knowledge points of the article, welcome to follow!

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.