Server Dictionary: Hot Standby FAQ

The so-called hot standby is to install the central server into two servers that are backed up each other, and only one server runs at the same time. When one of the servers running fails and cannot be started, the other backup server will automatically start and run automatically (usually about 2 minutes) to ensure the normal operation of the entire network system! The working mechanism of hot standby is actually to provide an automatic fault recovery capability for the central server of the entire network system. This article answers some common questions about hot standby.

Q: When do I need hot standby?

A: This problem is actually relatively simple. Generally, the server has to work for many years, and its backup work is absolutely indispensable. Therefore, to decide whether to use dual-system hot standby, first consider the importance of the system and the end user's tolerance for service interruption, and then decide whether to use dual-system hot standby. For example, how long can users in the network tolerate recovery services? What are the consequences if the service can't be recovered quickly?

Q: I have adopted RAID technology and data backup technology. Is it necessary to do hot standby?

A: There is no clear distinction between this, RAID and data backup are equally important! Data backup can only solve the recovery after a problem occurs in the system; and RAID technology can only solve the problem of the hard disk by experience. We know that when there is a problem with the server itself, whether it is the hardware problem of the device or the problem of the software system, the service will be interrupted, and the RAID and data backup technology can not solve the problem of avoiding service interruption. Therefore, for network systems with high security requirements and continuous and reliable application services, hot standby is still very important. In fact, we can think like this: If your server is broken, how much time do you have to restore it to work properly? This way you can understand the importance of hot standby!

Q: What is the difference between a hot standby solution and a cluster?

A: Conceptually, hot standby is one of the clusters. Clusters generally include two types: one is a pure application server cluster, that is, each application server accesses a unified database server, but some do not need file sharing storage, etc., such a cluster is relatively simple. The other is the dual-system hot standby of the database server. This dual-system hot standby implementation generally uses two shared servers at the same time, and in the general case, both take the primary and backup methods (also have high-end The system adopts a parallel approach, that is, two servers provide services at the same time).

Q: How do database services use hot standby?

A: The hot standby is implemented by software. That is, instead of using a shared storage device, the native data can flow directly between multiple hosts. Obviously, the biggest advantage of this method is that it saves expensive investment in storage equipment, and its shortcomings are not difficult to find: it will produce inconsistencies in data, or it will affect the speed of database reading. Let's look at an example: If you switch to the backup server when the service is interrupted, there may be a small number of transactions that have already been completed on the host that have not been implemented on the standby. Unlike the recovery of backup data, after the standby is started, the subsequent operations have already been performed, so it is quite difficult to retrieve the lost packets. This approach is therefore suitable for systems that are not very sensitive to the loss of small amounts of data.

Here is a standard solution for hot standby based on shared storage and dual-machine software. It provides fast switching in the absence of unattended, and there is no data loss, and investment in purchasing storage devices will be higher.

Q: How do I choose and implement a hot standby configuration?

A: 1. Take the application as the leading factor and conduct a serious analysis. Based on high availability.

2. Consider whether to use database hot standby or application server cluster or software backup.

3. Select to determine the specific device, software model, etc. The author here reminds everyone that different software or hard disk storage devices have compatibility problems between them. Therefore, you should consult a professional before purchasing, and do not appear to be incompatible with the related storage devices. phenomenon.

4. After the implementation is complete, be sure to test to ensure that the work is normal, and should pay attention to the regular test of whether the system can switch normally during the operation.