Oracle Dual-Machine hot-standby concept

1, dual-Machine Hot Preparation Overview
Dual-machine heat with two implementation modes, one is based on shared storage devices, the other is not shared storage devices, generally referred to as pure software mode.
Dual-Machine hot standby based on storage sharing is the most standard scheme for dual-machine hot standby.
In this way, two (or more) servers are used, using a shared storage device (a disk array enclosure or a storage Area network SAN). The two servers can be used in different ways, such as mutual preparation, master-slave, parallel. In the course of the work, the two servers will be provided with a virtual IP address to the external service, depending on the mode of work, the service request sent to one of the servers to assume. At the same time, the server detects the state of the other server through the heartbeat line, which is often done by establishing a private network. When one server fails, another server makes a decision based on the heartbeat detection and switches to take over the service. For the user, the process is fully automated and completed in a very short period of time, thus not impacting the business. Because of the use of shared storage devices, both servers are actually using the same data, which is managed by a dual-machine or cluster software.
In the case of pure software, the data can be copied to another server in real time by the two-computer software that supports mirroring, so that the same data exists on both servers, and if one server fails, it can switch to another server in time.
There is another case of pure software, where the server simply provides the application service and does not save the data (for example, only some calculations are used as an application server). In this case, it is also not necessary to use a shared storage device, but you can use either the dual-machine or the cluster software directly. However, this situation is not related to mirroring, but is only a small change in the standard dual-machine hot standby.

2, the database server dual-machine hot standby benefits
The advantage of this configuration mode is advantageous to the database upgrade, when the Systema need to upgrade, the service switch to run on SYSTEMB, upgrade a DB2 program, after the service can also switch back to a, and then upgrade the DB2 program B. This upgrade process does not affect the user's DB2 use, because there is always a machine that can use the DB2 program to respond to a user's service request.

Server failures can be caused by a variety of causes, such as device failure, operating system failure, software system failure, and so on. Generally speaking, it may take 10 minutes, hours or even days for the technician to recover the server in the case of the site. From a practical experience, it often takes more than a few hours to restart the server (which may still exist) unless it is simply restarted. And if the technician is out of the field, the time to restore the service is longer.

For some important systems, it is difficult for users to tolerate such a long service interruption. Therefore, it is necessary to use dual-machine hot standby to avoid long service interruption and ensure long-term and reliable service of the system.

Decide whether to use a dual-machine hot standby, the correct way is to analyze the importance of the system and the degree of tolerance for service interruption, so as to determine whether to use dual-machine hot standby. That is, how long your users can tolerate recovery services, and how much the impact will be if the service does not recover.

In considering the dual-machine hot standby, it is important to note that the general sense of dual-machine hot standby will have a switching process, the switching process may be about a minute or so. During the switchover process, the service is likely to be interrupted for a short time. However, when the switchover is complete, the service resumes normally. Therefore, the dual-machine hot standby is not seamless, uninterrupted, but it can ensure that in the event of a system failure, can quickly return to normal service, business is not affected. If there is no dual-machine hot standby, there may be a few hours of service interruption in the event of a server failure, and the impact on the business can be severe.

It is also important to stress that the server's failure differs from that of switches and storage devices, and its concept is much higher. The reason is that servers are much more complex than switches and storage devices, and they are complex systems that include both hardware and operating systems as well as application software systems. Not only is a device failure likely to cause service outages, but software problems can also cause the server to malfunction.

It should also be noted that some other protection measures such as disk array (RAID), data backup, although very important, but can not replace the role of dual-machine hot standby. The relationship between hot standby and data backup for two machines some users have the problem when planning a dual-machine or dual-machine backup: I already have raid, and tape backup, and do I have to do a dual machine? Or, if I do a two-machine backup, is it necessary to do a tape backup? It should be said that raid and data backup are important. However, RAID technology only solves the problem with the hard disk, and the backup only resolves the recovery after a system problem has occurred. And once the server itself problems, whether it is the hardware of the device or the problem of software system, will cause service interruption. Therefore, RAID and data backup technology do not solve the problem of avoiding service interruption. Dual-machine is also very important for systems that require continuous and reliable application services. Just think about it, if your server is broken, how much time you have to restore it to work, and how long your users can tolerate recovery time to understand the importance of a dual machine.

On the other hand, RAID and tape backup are also very much needed. For raid, the reliability of the system can be greatly improved at a very low cost, and its complexity is much lower than that of the dual machine. Because after all, the hard disk is the most mechanical operation, the most vulnerable parts, if the use of RAID, the failure of the system can be easily repaired, but also reduce the number of server downtime to switch. Data backup is an essential measure. It's a real-time backup, whether it's raid or dual-machine. Any software errors, virus effects, misoperation, etc., will be affected synchronously in multiple copies of the data. Therefore, be sure to make a backup of the data (whichever media you take, it is recommended that you have at least one offline backup) so that you can recover when data becomes corrupted and lost.

Oracle RAID differentiation using: http://www.cnblogs.com/wzmenjoy/archive/2011/12/21/2367642.html

Oracle Dual-Machine hot-standby concept