Operation and maintenance of three-blade flow

Objective

As a senior pirate fan, senior operations engineer, today I would like to do a fusion here, tell me about my three-dimensional operation.
This article I think everyone should guess, I am a sea thief fan. But you do not necessarily know my experience, I have in Jiangxi's most luxurious telecommunications room-Red Valley Beach Telecommunications Room (with the country's most luxurious room or some gap) work, personally involved in the building of the computer room, the network from scratch, the server from scratch, the team from scratch.
Here I have accumulated my original operation and maintenance of the concept and model, that is, three-knife flow operation model. After this after several moves, from the machine room operation and maintenance I stepped into the system software operation and maintenance work, from the bottom of the environment, into a higher level of software architecture maintenance. After that, I found that this model can be handy in today's operation.

Introduction to the Pirate King

Brief introduction
"One PIECE" (The Thief King, the Navigation King) is the manga work which the Japanese cartoonist Oda Eiichiro paints, this is a youthful blood inspirational cartoon.
Story background
A man with wealth, fame, power, everything in the world, "the Thief King." D., Roger, said a word before he was executed, so that all the people of the world were flocking to the sea. "Want my Treasure?" If you want, then go to the sea to find it, I put it all there. "The world begins to greet the era of the" Sea Thief ".
"The Time of the Sea Thief", in order to find the legend of the Pirates Wang Rojie left the big treasure "one PIECE", countless pirates raised banners, fighting each other. A teenager named Luffy to go out to sea with the promise of his severed arm for saving him, and to continue to seek like-minded companions on the journey, and began a great adventure to become a pirate king.
Three-Sword swordsman-Roroa Solon
Able to freely manipulate the three knives to fight. Love to drink, love to sleep, to talk about loyalty, Pirates of the first Super road fetish.
The road to the world's first swordsman was set for a childhood and close friend, and then became the protagonist BTs Mo Chit · D. Luffy's first partner. After losing the first sword to the world, Eagle eyed Hawk, he vowed never to lose again, and to exercise himself more diligently. Two years later he succeeded in meeting with his companions and, in order to realize his dream, went to the strong new world of the cloud.

First three-knife flow

In the computer room a lot of server management There are many things, but to classify things to see, the main thing is three types and can constitute a basic prototype of three knives.

• Monitor -and-word word
  Engine room operation and maintenance of monitoring is the most important thing, with the monitoring, we can see clearly now network, the computer room when there is a fault, can timely feedback to the customer fault information, fault analysis, let the fault in the first time to solve. At the same time with the monitoring we only good grasp the data, in the continuous analysis of data on the operation of the business to continuously optimize and improve. Monitoring is the best sharp knife to handle operational problems.
  
• Statistics -Three generations of ghost Beecher
  As the engine room operation, one of the main responsibilities is to manage the customer's assets, so clear the assets of the room is a very important thing, so can effectively statistics room assets, clearly know where the assets of the room, this is very key. And in the previous company statistics every day to come to trouble, very headache. But a good grasp of statistics can quickly deal with problems to improve efficiency, save time to find. Statistics this knife to use well very good, but with bad oneself to make oneself headache, because the statistics out of the thing is unclear, so with demon knife three generations of ghost Beecher to describe should be more appropriate.
  
• Maintenance -Black knife eyes
  Maintenance is mainly to maintain the methodology, each maintenance of the way to organize, step, standardize, process, tool, service, and ultimately achieve the purpose of machine management machine. Through the collection and collation of methodologies, steps, standardization, process, tooling, and service, operation and maintenance efficiency will be a geometric order of magnitude of ascension, so here with the big sharp knife-black sword eyes to describe a little too.
  Three-knife flow kind in the engine room operation of the three-blade flow has a first effect, then I entered the software system operation and Maintenance field, the first into the company, I am also very curious about the three-knife flow before the application, with a disturbed mood, I entered the post. Perhaps due to the similar business model, I found that my three-blade operation is still useful, but also very effective, although not able to deal with the problem immediately, but the principle of basic similarities, to solve and deal with the daily operation of the problem is very simple. All I had to do was expand the range and I began to hone my knife.
  
• Increased monitoring range
  Previously in the computer room monitoring equipment, only monitor the flow of the use of the basic can meet the daily needs, because the room server numerous, to monitor all the servers have and only monitor traffic. But in the application maintenance in addition to monitor the flow of CPU, memory, SQL query and so on, will cause the system card slow all the performance indicators are best to monitor.
• Widening of the statistical range
  Statistics are not just statistical assets so simple, from business use, load situation, knowledge management and so all need statistical records, and when the statistics recorded to a certain amount of time, we will find a lot of things are interlinked. By statistics we can analyze the object of the matter and how to solve the problem. This time we need to use the appropriate management tools to help us to count and analyze things. Fortunately, I met a knowledge management software, Wecenter, this software can provide the type of question and answer management. Very good, everyone is interested to understand the trial.
• Maintenance Mode Upgrade
  Maintenance mode before in the room just walked through the steps, flow, into the new operation and maintenance environment, I began the tool (programmatic) work, and towards the service of progress. I started using jerky programming techniques to write gadgets.
  Through these steps, I have basically completed the efficient handling of daily affairs. In particular, the system configuration from the original manual step-by-click Configuration, about half an hour to an hour of the process gradually shortened to 5 minutes, the configuration steps down to 10 steps, greatly reducing the configuration complexity. Introduction of three-knife flow in a knife flow of the lion song monitoring through monitoring, we can collect the system at different times of the operating data, aggregated together for graphical analysis, we can speculate on the situation of the problem, to restore the cause of the matter, and the results.
  This trick is almost a kill in a hit, because any problem will leave traces, and these traces will be displayed in the chart through the data, through monitoring records, analysis of the nature of the restoration of things. So that the problem is fundamentally dealt with.
  As an example:

  After an update, the system often appears to die, this kind of card-like situation always can not find the problem where, previously considered a database problem, kept in the database to do optimization, but in the end all the optimized indexes are optimized, the problem remains.
  This time the surveillance, Through the analysis of the monitoring data found that the Web server each time the card is dead, there are several indicators of abnormal elevation 1. Is the number of processes, 2. Is the number of TCP connections, after the TCP packet, found a large number of empty connection requests to the database, and later put this issue into a document submitted to research and development this thing has been low-key

36 Troubles Phoenix Statistics

The statistics here refer to asset statistics, project statistics, and configuration statistics. With these statistics, we can manage the content of the whole system very accurately, we know what to do, and we can clearly understand what we are going to do, what we should do, because all the assets can be seen, at a glance.
Statistics is a very burning thing, because a change will be all changed, supporting the action to be processed in place, in the computer room when used or execl table so processing a lot of data statistics is very laborious, often error, the result is often injured himself.
But statistics do make work more efficient, for instance.

A complete table with the machine configuration information, location information, System information, in a table can be found all the information.
One thing is distributed, as long as the machine records, maintenance needs of the data are out, where is the machine? What configuration? What system? You can work immediately. And everyone knows all the questions.

One-knife flow • 360 troubles Phoenix Maintenance

The maintenance mode is arranged, the maintenance mode is step, the process is formed, and the standard template system is established. Once this template system is established, we can try to program, automate, and combine it with the workflow. The following is a practice that I have in step, process, and form a standard template system.

Engine room operation and maintenance of the process of doing the wrong system I think is the most headache of one thing, I believe you have encountered or heard of the computer room to do the wrong system to clear the data, to the end of the computer room to recover data loss.
No one wants to make mistakes, but err. The key is how to prevent and avoid mistakes.
In order to solve the above problem, the leadership taught me a best practice-each time to do the system to take a pen to record things in the book, in the Machine room operation.
After this method, I self-designed a computer room things statistics record table, and print out, form a table, which solves the problem often to be counted two times, the computer statistics once, in the book once, to prevent the small partners in order to lazy only on the computer memory and then by memorizing operation caused the problem of error.
Mandatory requirements of small partners to notify the things must be recorded in the printed form, after several checks clear, see clearly in the implementation, in order to avoid the omission of the wrong memory error occurred in the probability.

After a period of time the effect is very good, to do things more clear and easy to organize. It turns out that a good process-based approach to design can avoid errors and improve operational efficiency. Solve the challenges in operations.

Three-sword flow-a big · 3,000 • the world

Before the Solong chant a few lines: "Nine mountain eight sea, for a world, poly-Thousands of boundaries into a" small thousand world ", this sector by three, without my misconduct, three knives flow of secret .... A big one? 3,000? the world!

I quote this passage is to illustrate my point of view, the future of operations, distributed management, will be less poly, sand into the mound, through the operation of three knife conditioning, I want to achieve tens of thousands of server management will not be difficult. Of course this is an ideal condition for me.

End:

As an OPS worker, I think the data in the operation and maintenance work in the collation of learning and analysis should occupy 80% of the work and operation should be in the overall workload of 20% of the daily collation of learning and analysis for the operation of the service.
At present, our operation and maintenance of the center of gravity is just the opposite, Operation first, learning summary second, and the operation and maintenance of the evaluation indicators with a large number of factory piece of the way to consider operations, operations into a system for the operation, but most bosses will like to do so, because they pay attention to the results, They pay attention to the result that you took the money to do for me how much I need you to do the work, other things you do not need to consider and do not consider.
Do not consider too many problems in the work is correct, but the result is that everyone is the boss of a tool to make money, mutual adhesion is poor, some deep problems can not be real feedback, over time problems accumulate more tired, the end will go to the brink of collapse.
For operations, a large number of operating time to squeeze out the summary of the time to study thinking is disastrous consequences, lack of protection, experience is not cumulative (can not properly record the problem, no time, can only take care of their own), everyone is repeating the pit, repeat the wheel, a matter asked several times, asked the last no friends, resulting in product quality decline, problems, repeated repair problems, work often make mistakes.
As the saying goes, often in the river where there are not wet shoes, often in the lakes and rivers where there is no knife, those who did not miss the matter of "sage" must have not done the operation of the dimension. It is easy to cause instability when the system changes frequently, and big changes mean big problems, big risks. Every change means stepping on a hole and making a mistake. How to not fall into the pit is the problem that every operation has to face.

"Change is the norm, and it is necessary to establish a control change under the normal mechanism to give a possible fault"

System change is a normal operation, is to operate the cost of eating, the system in the iteration, to meet the needs of the need to change, as the OPS is to eat the operation of this bowl of rice, will not be able to unworthy.
So I want to stabilize the system, I think the method is to reduce the implementation of human change, the scope of the impact, the introduction of change control, control the impact surface. These are practical and detailed cases in the book on continuous delivery-a systematic approach to the release of reliable software.
Operations are really doing things, should be constantly found to solve the problem of the matter, OPS should have the spirit should be like Roroa Solon, want to when number one big sword Hao, like a thief King promised to never lose the same. Constantly to improve themselves, enrich themselves, challenge themselves and surpass themselves.

Light up your big swords! The brothers of operation and maintenance! My friends! Cut it out for the raging problem! When you cut off all the problems you will have a sense of pride and glory, and the goal of being a big sword will be in the corner.
Finally want to say a word, please respect each friend around, meet friends and acquaintances are the fate, the fate may not have a part, the margin is due to the fruit, may all friends knot good because of good fruit.
Finally thank you for the product read, the article mechanically in the place also please Haihan

?

Operation and maintenance of three-blade flow