Five indicators for MySQL database O & M

How to evaluate a company's database O & M level? What is used for horizontal and vertical comparison? What is the goal of automated platform construction? The corresponding indicator system must be provided for guidance and must meet the following conditions:

• It can be measured and measured using numbers.
• Final indicator, not intermediate indicator

For example, DBA may pay attention to the database throughput, but the higher the throughput, the better the service quality provided by the database, the reason why developers are concerned about this indicator is also because they are worried that a high throughput will affect the response time or cause the system to be unavailable. Therefore, this is only an intermediate indicator.

• The database O & M level of a website can be fully measured without losing sight of it.
• Humanistic Attention

Recommended reading:

Production Environment MySQL master/Master synchronization primary key conflict handling

MySQL + KeepAlived + LVS single-point write master/Master synchronization high-availability architecture Experiment

MySQL master/Master synchronization Configuration

MySQL master-slave copy notes in CentOS 6.3

1. Data Security

Data security is the first priority. DBAs must ensure that data is not lost and data is lost!

This has three meanings:

1) when a human error occurs (update, insert, delete, drop, alter), the data can be restored to the correct state.

2) data can be restored to the correct status when data centers, hardware faults, operating systems, and database software are faulty.

3) No transactions are lost to ensure that the data that has been written into the database can be correctly queried.

In addition, you must ensure the consistency of the master and slave databases. Otherwise, data is still lost in the case of read/write splitting.

For 1, backup is mainly used to ensure that, because replication can be used for disaster tolerance, it cannot be used for Fault Tolerance (of course, delayed backup is acceptable to a certain extent ).

For 2, backup may be used for recovery, or the master database or slave database may be switched directly to restore the service.

For 3, e-commerce, the payment Library requires a very high level, using the highest security level of Database hardware and software settings and redundant equipment, the goal is not to lose any one transaction, this is because even one transaction may cause a lot of money loss and a decline in the company's reputation.

The "911" incident caused 1200 companies to be affected. More than half of the enterprises were declared to collapse because IT data was damaged and lost, leading to business failures. Financial giant Morgan Stanley's global business department recovered to normal the next day because the previously established remote disaster recovery system protected important data.

Measurable indicators:

• RPO (Recovery Point Object): the Recovery Point indicator refers to the data at which Point in time the disaster Recovery system can recover data after a disaster, it measures how much production data is lost after a disaster
• RTO (Recovery Time Object): System Recovery Time

RPO describes the reliability and integrity of backups, and RTO describes the reliability and speed of recovery.

Because the open-source MySQL version does not provide hot backup tools and Backup management tools (MSSQL, Oracle, and of course commercial software ), therefore, the DBA is required to develop its own Backup Recovery Management Platform (SCRIPT ). This is also the first job of DBA.

For more details, refer to page 2nd:

1 2 3 4 Next Page

[Content navigation]
Page 1: Data Security	Page 1: No fault (downtime)
Page 1: Response Time	Page 1: Cost