What does database O & M do?

I.Database Lifecycle

The database lifecycle can be roughly divided into several stages based on the software lifecycle and project development.

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131228/22462324P-0.png "title =" 1.png" data-pinit = "registered"/>

1.Planning

After the project is established, the software and hardware selection of the database platform and the general database architecture are selected.

1.1 configure the number of servers, memory size, disk space, IOPS, CPU cores, and network bandwidth of the servers;

1.2 The OS and database products selected and the corresponding version;

1.3 overall architecture, such as whether to consider policies such as HA, Scale out, load balance, and read/write splitting.

2.Development

Development work is usually carried out in the test environment. After development, it is moved to the production environment.

2.1 database design;

2.2 SQL programming and debugging;

2.3 SQL optimization during development.

3.Implementation

Deploy the developed database program to the production environment. Now, the project is launched. The O & M phase will be followed.

II.What O & M do

As shown in the figure above, O & M is the work after the project is launched. Let's see what O & M has done since the project was launched.

1.Deployment Environment

1.1 if there are too many servers installed on the database, you can choose Silent Installation );

1.2 parameter configuration of instance and database parameters );

1.3 grant logon and Database User Permissions ).

2.Backup/Restore

For databases, it is very important to have an available backup to prevent data loss caused by data corruption and user misoperations. O & M makes sense to ensure the existence of data, Otherwise other work will be useless.

3.Monitoring

For O & M, we must first ensure the operation of the database, and then the performance of the running system. Therefore, monitoring is mainly divided into the following two points:

3.1 is there any database interruption or exception, error or warning in the database running status?

3.2 database performance. Are there any performance problems or performance risks?

4.Troubleshooting

Diagnose and fix database errors or warnings reported by system users during monitoring.

5.Performance Optimization

The database performance problems discovered during monitoring or reported by the system users are optimized.

6.Disaster Tolerance

Disaster Tolerance is only a means. In the end, to ensure the availability of the system, the following policies are usually selected: failover cluster, image, log transmission, remote backup, and so on.

If you have deployed a disaster recovery policy during implementation, you only need to perform status monitoring.

The disaster recovery policy is also added after the system is launched for a period of time.

7.Upgrade/Migration

7.1 upgrade

It is usually performed on the local machine and the hardware remains unchanged, for example, replacing the operating system, database version, and patching;

7.2 migration

Generally, hardware needs to be upgraded. For example, if a new server is replaced, the database is moved to the new server;

There is also a "Migration" on the local machine, just to move the database file location.

7.3 migration + upgrade

However, most of the time, they are upgrading during migration, that is, changing the server or software version.

8.Health Check

It is usually called inspection or HealthCheck. It may be daily, monthly, or yearly.

In fact, if the inspection content is daily, hourly, or even every X minutes, it is a quasi-real-time system monitoring.

9.Database problems reported by system users

DBAs are required to handle any database problems reported by users, even if it is not a database problem.

In a broad sense, all tasks except database development should be considered within the O & M responsibilities.

Q:What are the daily tasks of database O & M?

A:You can list all the tasks to be done one by one.

For example, database running status monitoring includes: database service interruption, disk space, Error Log check, database consistency check, job running status, index fragmentation check, and so on.

The detailed list of each task will be broken down one by one.

3..Problem solving during O & M

If you encounter problems during O & M, it is good to solve them through your own/others' experience;

However, if there is no solution, it is usually as follows:

1. Check logs: Are there any related errors/Information prompts in the operating system/database/application logs;

2. Check the error code: Is there a solution in the official document/shared by netizens;

3. If none of them are found, they will win the prize. If they do not, they will analyze the team and the team will not find the official support. Of course, sometimes the official support may not be solved.

Note:For online systems, such slow query may take too long and affect user experience. It is usually a matter of priority to quickly solve the problem. The only fear is to use a temporary emergency solution to ensure the availability of the system, and then analyze the root cause and solve it thoroughly to prevent the next occurrence.

This article is from the "SQL Server DBA" blog, please be sure to keep this source http://qianzhang.blog.51cto.com/317608/1198503