Redundancy and hot standby

Thread Content

Redundancy refers to repeatedly configuring some components of the system. When the system fails, the redundant components step in and assume the work of the failed component, thereby reducing the system's failure time. Redundan, automatic backup, that is, when a certain device is damaged, it can automatically replace the device as a backup device. The main redundant system accessories include: Power supply: Dual power supply systems are commonly used in high-end server products. The two power supplies are load balanced, that is, they both provide power for the system when the system is working. When one power supply fails, the other power supply takes on all the load. Some server systems realize DC redundancy, and other server products such as Micron's NetFRAME 9000 realize full redundancy of AC and DC. Storage subsystem: The storage subsystem is the most prone to failure in the entire server system. There are several ways to achieve redundancy in this subsystem. Disk mirroring: Write the same data to two disks respectively: Disk duplex: Adding an I/O controller to the mirrored disk creates a disk duplex, which improves bus contention; RAID: Abbreviation for Redundant array of inexpensive disks. As the name suggests, it consists of several disks, and a single data stream is written to these disks in sequence through a controller coordinating motion mechanism. The RAID3 system consists of 5 disks, of which 4 disks store data and 1 disk stores verification information. If one disk fails, the failed disk can be replaced online, and the data on the new disk can be recreated using the other three disks and the verification disk. RAID5 distributes the parity information on 5 disks, so that any disk can be replaced, and the rest is the same as RAID3. I/O card: For servers, it mainly refers to the redundancy of network cards and hard disk control cards. Network card redundancy is to plug in dual network cards in the server. Redundant network card technology was originally a technology found on mainframes and mid-range computers, and is now gradually owned by PC servers. PC servers such as Micron's NetFRAME9200 can achieve redundancy of up to four network cards, and each of these four network cards bears 25% of the network traffic. All Compaq's ProSignia/Proliant servers have fault-tolerant redundant dual network cards. PCI bus: NetFRAME 9200, a product representing the highest technical level of Micron, uses triple peer PCI technology to optimize the bandwidth of the PCI bus and increase the data transmission speed of high-speed devices such as hard drives and network cards. CPU: The main processor in the system does not often fail, but symmetric multiprocessors (SMP) allow multiple CPUs to share work to provide some degree of fault tolerance. The so-called dual-machine hot backup is to use two servers that back up each other to perform the same service. One host is the working machine (Primary Server) and the other host is the backup machine (Standby Server). When the system is normal, the working machine provides services for the application system, and the backup machine monitors the operation of the working machine (the working machine also detects whether the backup machine is normal). When the working machine is abnormal and cannot support the operation of the application system, the backup machine actively takes over the work of the working machine and continues to support key application services to ensure uninterrupted operation of the system. Under what circumstances is it necessary to use dual-machine hot backup? Users can decide whether to use dual-machine hot backup based on the importance of the system and the end user's tolerance for service interruptions. For example, how long can users on the network tolerate for service restoration? What will be the consequences if the service cannot be restored soon? Servers responsible for key enterprise business applications require extremely high stability and availability, and require 24/7 uninterrupted service. It is recommended to use dual-machine hot backup. I already have RAID and tape backup, do I still need to have two machines? Or, if I do a dual-machine backup, is it necessary to do a tape backup? RAID and data backup are both very important. However, RAID technology can only solve hard disk problems, and backup can only solve system recovery after problems occur. Once there is a problem with the server itself, whether it is a hardware problem or a software system problem, it will cause service interruption. Therefore, RAID and data backup technology cannot avoid service interruptions. For systems that need to continuously and reliably provide application services, dual machines are still very necessary. Data backup is an essential measure to ensure data security. Because no matter RAID or dual-machine, it is a real-time backup. Any software errors, virus effects, misoperations, etc. will affect multiple copies of data simultaneously. Therefore, even if a dual-machine solution is used for key businesses, data must be backed up so that it can be restored when the data is damaged or lost. How to choose and implement a dual-machine hot standby configuration solution? 1. Be application-oriented and high-availability as the purpose, and analyze the necessity of requirements; 2. Select and determine specific equipment and software models, etc. It should be noted that there are compatibility issues between different software or storage devices such as hard disks. Therefore, you should consult professionals before purchasing to avoid incompatibility with related storage devices after purchasing dual-machine software. 3. After the implementation is completed, testing must be carried out to ensure that it is working properly, and attention should be paid to regularly testing whether the system can switch normally during operation.