Can't open the game, saying that the server is under maintenance. What do you mean?

Come on, as a former Netease game practitioner, let me talk about what I am doing in the real server maintenance process.

There are two kinds of server maintenance, emergency maintenance and routine maintenance.

1, emergency maintenance

Emergency maintenance is generally a hardware failure or a serious Bug. This period of time is the most tense time for each team. Every team is busy.

The operation team will issue announcements, comfort players, count losses, prepare fault reports, evaluate fault levels and so on.

The planning team may consider the impact of bugs and decide whether to return the documents or what to compensate, whether the compensation will affect the economic system, and so on.

The most important thing for the program group is to reduce the influence of bugs, such as hot updating the shielding gameplay and then solving bugs. Sometimes hardware failure will lead to unbalanced load on some nodes of the whole server, or in case the database server dies, you have to cut the database quickly.

The QC team will continue to test it, including the revised version. QC boss usually has a veto to decide whether to release a version.

SA team will have someone on duty to follow closely. If there is something wrong with the hardware, contact and change it or something.

2. Daily maintenance

Daily maintenance is much more streamlined.

The operation team will release the updated content in advance to warm up the new gameplay.

After the planning meeting, I will run the game on my computer to see if there are any problems, and then go to the forum to see the players' reaction after the official opening of the service.

The general procedure is to update the code, deploy it to the server for testing, and if the architecture changes, use SA to run the script to change the architecture. At the same time, another group of people will run a bunch of scripts, including bills reconciliation scripts, scripts for importing data, scripts for updating various lists, scripts for preheating databases and caches, and so on. Everyone will also test their own modules and then stare at various alarms.

The QC team will also test this week's new gameplay and main process on the server again.

Other support teams work together every day.

In addition, I saw a lot of replies saying that I would restart the server. In fact, I want to say that we really rarely restart the server. The cost of each restart is very high. The server will not restart unless the hardware is updated or under special circumstances.

In fact, generally speaking, server restart can be understood as two kinds, one is hardware restart and the other is server process restart. Hardware restart is rare, and server process restart is basically necessary.

Many people ask why hardware restart is so expensive. Feel your answers. Let me just add a comment.

If it is an automatic startup program, there is no problem. However, many system configurations and software settings of the server must be reset after startup. At the same time, we have to communicate with other servers to restore the state before startup, or synchronize the information of other servers to correctly handle various operations during shutdown. There are many things, heavy tasks, high probability of mistakes, and reluctance to restart.

Others asked why it lasted so long. I want to say, please post in the customer service area or forum, and a large wave of operators will come to appease you.

Others complain about maintenance twice a week. I want to say that Eve's masterpiece is to maintain it every morning. Maintenance is also for everyone to play better. (Spit out who let EVE use a Windows server. )