Shulou(Shulou.com)06/02 Report--

This article mainly introduces the relevant knowledge of the problems in the operation and maintenance development, the content is detailed and easy to understand, the operation is simple and fast, and has a certain reference value. I believe you will gain something after reading this article on the problems in the operation and maintenance development. let's take a look.

As long as operation and maintenance development is involved in the operation and maintenance work, there are bound to be several problems to be solved in order to promote this matter:

Raise the awareness of operation and maintenance. Work should be done from bottom to top and from top to bottom, the value and gold content of the upper operation and maintenance work can be recognized, and our work can improve efficiency and liberate ourselves. For example, for operation and maintenance development, I can cooperate and coordinate, and there are technical difficulties that can be solved, but I will not chase others to learn certain technologies, because this kind of thing will change, and there is this in the consciousness of operation and maintenance. Then the meaning of this matter is very different.

There should be clear objectives for operation and maintenance. It is clear here that planning alone is not enough, and it is necessary to have a clear goal of operation and maintenance. from another perspective, this goal is the pain point of our work, and solving the pain point of work is the promotion of our own consciousness. in this way, it can also explain the significance of achieving the goal of operation and maintenance.

There should be a clear time window. With goals, we need to arrange a designated time window to do, if there is no time frame, then the progress and quality of things will be difficult to trace and guarantee.

I have stumbled a lot in this matter, and I will find that things are getting more and more out of control. For example, my expectation is 6, and the result is 2. The greater the contrast, I find that there is so much room for improvement that I will fall into an endless cycle. I will come up with a lot of improvement methods and suggestions. But these methods and suggestions will be abstracted into a series of improvement tasks, which involve front-end, back-end and design, so I need to confirm, communicate and implement each point. Then the progress of things slows down. I still remember this sentence when dealing with the operation and maintenance platform with the same execution efficiency as "Mad Dog." sometimes I would ask me why I insist on doing this thing, for what purpose, and what good it is for us. It's easy to give up, and this sentence supports me: when you want to give up, think about why you started in the first place.

Let's get to the point, that is, whether a "tragic" deployment and installation scenario is tragic, how to solve those problems, and how it was entangled at that time, you can listen to my thoughts.

First, let's talk about the scenario of instance deployment.

Suppose the following is a preliminary installation and deployment page.

There are some goals that can be achieved to achieve this function. For example, we can now implement page call script content, let's take a look at what needs to be noted, or easy to make people tangled.

First of all, it may not be easy to understand whether this requirement meets expectations. For example, if our requirement is to deploy a set of database software, then whether kernel parameters need to be adjusted, system parameters should be initialized for unified configuration, database additional plug-ins need to be installed, backup should be configured, monitoring should be deployed, and metadata should be generated automatically. Another point is whether there is already an example of this environment, if so, then the default configuration of / etc/my.cnf needs to be readjusted, so this seemingly simple page will not meet the requirements, so we expand and converge. The above functions are all basic requirements, which we need to consider, but not all the details need to be executed uniformly, for example, kernel parameter optimization may be initialized once. Therefore, we need to refine the scripted work and be able to achieve modular functions, which involves some logical changes and optimization.

Of course, from the front end, a difficulty is that the log implementation results back, can basically achieve the effect of refresh. For many operators, this requirement is relatively difficult to achieve. Therefore, it can be abstracted into two difficulties, one is the progress display, the other is the log display, in which the log display is the most important.

It seems that the scripting work is almost done, assuming that we have spent some effort to make customizations and changes. Then the next thing is the execution of the script. I still have to quote a picture I drew before.

For example, if we want to deploy the environment, the execution path may be ops (operation and maintenance platform)-> CM (central control)-> DB Server (server), or ops (operation and maintenance platform)-> DB Server (server). For example, from a standardized point of view, ops (operation and maintenance platform)-> CM (central control)-> DB Server (server) is a more appropriate path, then the script is triggered from the OPS side, how to reach the DB Server side Because there needs to be a transit process, you can use paramiko,ansible or salt to complete it, but how to connect seamlessly is a difficulty. From the point of view of access management, it may be abstracted into an interface. It seems to be a matter of command, but it is really troublesome to link up.

Then let's talk about the requirement of data query, which is actually a very basic requirement. The data can be displayed through the interface to meet certain query requirements. But there are a lot of problems and uncertain needs.

For example, if I want to implement data query, according to the above diagram, the execution path may be ops- > CM- > Server- > DB, this path is very long, or ops- > CM- > DB. If this path is selected, how to activate permissions is a difficult problem. We assume that there are 100 MySQL servers, write a script to batch transfer scripts to these 100 servers, create a user on the database, and then activate firewall permissions. It seems to be a very simple thing, but you will find that this method is wrong, for example, there is master-slave replication in it. If you execute in the slave library, the master-slave replication is likely to be disconnected, so you need to deal with more replication problems. So there are 100 servers, and you need to do a layer of extraction, that is, to find out which are master libraries and which are slave libraries, but you will soon find that judging master and slave libraries is still a hassle. Because the MySQL level does not have a clear identification of master or slave, but needs to be obtained indirectly. It is not intuitive to get the information of slave directly from the main database. If I confirm it one by one and feel that this way is too low, I will encounter layer after layer of problems if I want to link it up perfectly. The most embarrassing thing is that the metadata is not accurate enough. I am busy and find that I still lack some data.

Of course, we can struggle, or we can make improvements, and we can clearly sort out the boundary. For example, if we are dealing with the deployment of operation and maintenance, then we will focus on this place to see how much manpower and time costs need to be invested to solve the problem. A preliminary solution, can quickly iterate out some results. On the contrary, it will be found that it is a little iteration, but it is all improving and there is no shaping result.

This is the end of this article on "what are the problems in operation and maintenance development"? thank you for reading! I believe you all have a certain understanding of the knowledge of "what are the problems in operation and maintenance development". If you still want to learn more knowledge, you are welcome to follow the industry information channel.

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.