Avoid the misunderstanding of Disaster Recovery Backup

With the development of the information industry, disaster recovery and backup have been accepted by many companies and enterprises. At the same time, the manufacturers of disaster recovery and backup also threw out a lot of new concepts and ideas. However, as a professional disaster tolerance backup provider in the industry, Beijing Heli Kee Technology Co., Ltd. (hereinafter referred to as "Heli Kee") found that there are some misunderstandings in the industry, some ideas are too misleading, so it is necessary to clarify them one by one, which is conducive to the development of the industry and to the customer's choice of products.

Misunderstanding 1: Disaster Recovery Backup = copy

In the disaster tolerance backup field, some people think that the so-called disaster tolerance backup is simply copying files. This is a very superficial and wrong understanding! In fact, copying is only a regular backup, a cold backup, and a time window. For example, a hospital copies data at every morning, if a fault occurs at three o'clock P.M., data in this period of time (from three o'clock P.M. to, within six hours) will be lost, resulting in loss. At the same time, it is obviously not suitable to simply copy structured data such as databases, because it not only has data files, but also log files, and sometimes may miss copies, at this time, the logical relationship of the data may not be copied, resulting in unavailability of data recovery. In addition, in the big data era, copying a large volume of data also requires a lot of resources and time, which is prone to errors in the middle and can also cause business interruptions, this is intolerable, and it is a huge cost in some enterprises.

Therefore, you need real-time backup to achieve zero data loss. At the same time, you need to be familiar with the database structure and incrementally back up the database data to save resources and time, incremental backup is also required for unstructured data such as files, which can reduce the usage of bandwidth and resources.

Misunderstanding 2: image, snapshot = CDP

In the Disaster Recovery and backup industry, many backup services are backed up using image or snapshot technology, which is also known as CDP Technology (continuous data protection). This is totally wrong. Because according to the strict definition of CDP by snia (Global Network Storage Industry Association): (1) Real-time Data Backup; (2) remote backup; (3) backup data can be rolled back at any time point. Obviously, images and snapshots only meet the requirements of point (1), and the requirements of point (2) may barely meet, but the requirements of point (3) cannot be met.

The image data is exactly the same as the original data. It is similar to looking in the mirror and does not record the data change process. Therefore, it cannot be rolled back, and it cannot be rolled back at any point.

Likewise, a snapshot is a collection of multiple images. Similar to taking a photo, although it records the data change process, it can be rolled back because it is a limited record point, however, we cannot roll back any point. In addition, snapshots often have a lot of redundant data, which greatly wastes storage resources.

Misunderstanding 3: Dual-host = application-level disaster tolerance

Disaster Recovery Backup can be divided into data-level disaster recovery and application-level disaster recovery. Data backup only protects data. Even if a loss or logical error occurs, it can be quickly restored or repaired. Application-level disaster recovery not only ensures data security, but also ensures that the system does not stop and business systems are guaranteed.

A dual-host machine combines two machines and uses SHARED disk storage to automatically take over tasks when one machine fails. Obviously, this is only a solution of system redundancy and high availability (HA). Because there is only one shared data, once a storage problem occurs, not only data will be lost, but the system will also go down. Obviously, this is not a disaster tolerance backup solution, let alone application-level disaster tolerance.

Misunderstanding 4: the use of database backup tools can meet the requirements of Disaster Recovery Backup

Some people think that using the backup tools provided by the database, such as rman, mongouard, and goldengate, the backup tools provided by these database systems can basically implement local or remote backup of the database, however, they almost all implement data backup by means of preset time points or backup interval. The following problems cannot be solved: (1) Real-time Data Backup cannot be implemented, and the backup interval data is not protected; (2) during backup, the data is copied in a centralized manner over a period of time, high pressure on servers, networks, CPUs, etc. Most of them need to stop external services during backup, which affects business continuity. (3) it is difficult to use, non-DBAs are often not used.

Misunderstanding 5: we have recovered the backup data.

Some vendors guarantee that customers' data can be backed up in real time during product promotion, but do not guarantee that the data backed up in real time can be recovered (this is often not publicized ). In fact, the fundamental purpose of data backup is to recover an unrecoverable backup, which is meaningless to any system.

As an end customer, we must be aware that restoring data securely, conveniently, and efficiently is the real life of the backup system. Many may think that since the backup system has backed up the required data, the restoration should not be a problem. In fact, examples of data loss may occur from time to time, whether in a data center in the financial and telecommunications industry or in a common desktop system, backup data cannot be recovered.

Misunderstanding 6: databases are foreign backup software, and foreign backup software should also be selected

For various reasons, foreign database systems have a high market share in China. Therefore, many people think that since the database is foreign, backup software must also be foreign.

This is completely an error. Because the Disaster Recovery Backup system is very special, once a disaster recovery backup system is built, all data and even applications must be handed over to the Disaster Recovery Backup Software. That is, all data and applications are "Transparent" in the face of disaster recovery backup software ". Therefore, it is best to use domestic software for disaster tolerance backup systems. Otherwise, data security is nothing to worry about.

Hutchison is a recommended disaster recovery backup system. It can back up data (including databases and files) to the backup machine in real time and incrementally. Unlike the image and snapshot modes, it can continuously record the data change process, similar to a video recorder and performing any point rollback, this is the true CDP (true CDP ). Different from the dual-Host Architecture, the standby mode can not only back up data to a local or remote location, but also have a management module to implement dual-host functionality, ensure business continuity, and achieve application-level disaster tolerance. It is easy to use and can completely replace the built-in database tools, and save the DBA labor costs. It is even more worth mentioning that standby tejia is fully self-developed and supports the Chinese operating system-youbangtuo (Kirin Ubuntu, also known as ubuntukylin ), at the same time, it also supports the three major domestic databases-Human Resource Management, Shentong (Universal China) and dameng database, which can be fully localized to create a trusted secure disaster recovery backup system.

This article from the "disaster tolerance backup-Hutchison Yi Liu nazhe" blog, please be sure to keep this source http://471014.blog.51cto.com/461014/1534123