RAID technology 4: raid q &

What problems should individual users pay attention to when building RAID, or a disk array?
1. Q: What RAID solution should I choose, and the motherboard With RAID function? RAID Controller? Or software RAID?
A: In fact, RAID solutions are only divided into high-end and low-end solutions. For the vast majority of low-cost RAID solutions, their architecture does not include the computing part. Therefore, they are highly dependent on the CPU, low-speed CPUs are difficult to do this kind of job. Of course, for newer CPUs such, however, to ensure the stable operation of the RAID system and avoid dragging down the system performance by RAID, we strongly recommend that you use a CPU with a clock speed above 1 GHz.
Therefore, the choice of RAID card or motherboard With Integrated RAID function should be based on the user's needs. Generally, the use of RAID card can achieve relatively stable performance, however, it will occupy a valuable expansion slot, and the cost is high. If you are preparing to upgrade the motherboard or new installation, the motherboard integrated with the RAID chip is the first choice to implement the RAID function at the lowest cost.
2. Q: I used the RAID system, but I didn't feel a significant improvement in speed. Why?
A: There are several criteria for RAID systems. For RAID 1, RAID 5, and other standard disk arrays, data reliability is primarily pursued. Therefore, although it is a parallel storage, however, because data needs to be verified, their write performance will be affected. For common users, the speed improvement is not obvious. Of course, for some applications that require a large number of reads, their advantages are relatively large.
3. Q: I have used the RAID 0 standard disk array mode. I heard this will cause very insecure data. Should I use the RAID 0 + 1 mode?
A: The security of RAID 0 is acceptable for common users. However, RAID 0 is obviously a dangerous method for important data, in addition, the larger the number of disks in the RAID 0 array, the higher the chance of problems. Because RAID 0 uses data splitting without redundancy, once a disk fails, it will have a devastating impact on all data, redundant disks (or data blocks) are provided for backup or data recovery in other RAID modes. Therefore, we recommend that you do not
0 array stores important data or installs the system on it, because the storage and exchange of a large number of temporary swap files is the true advantage of RAID 0.
4. Q: I have two hard disks with different specifications. Can I use RAID? What is the effect?
A: You can use RAID 0, but note that the total capacity of the RAID 0 created in this way will be a smaller disk capacity × the total number of disks, as a result, some resources may be wasted. Of course, if the computer has strong computing power and uses Windows 2000/XP operating system, using software RAID can avoid this loss. For details, see the "Dynamic Disk" introduction in this article.
5. Q: What is the difference between using RAID and using SCSI hard disks? Which one is more cost-effective?
A: Here we only discuss the IDE-RAID. In general, SCSI can provide better stability and faster speed, but the price is quite expensive. The price of a 7200-to-9.2 GB/Ultra160 SCSI hard drive will be as high as about 1500 yuan. In contrast, the general IDE hard drive capacity at the same price will be more than 80 GB.
We can compare IDE-RAID and scsi raid, we use four Maxtor golden diamond seven generation 80 GB hard drive consisting of RAID 0 + 1, choose to support ATA/133 HighPoint RocketRAID 404 RAID card, the total price of this system is less than 8000 RMB *, which is equivalent to a 10 000 to GB SCSI hard drive, using Ultra160
SCSI interface/4 MB cache, and equipped with scsicards supporting Ultra160.
Of course, SCSI hard disks have many incomparable advantages, such as high transmission rate, low CPU usage, and support for hot swapping. However, for individual users, their cost effectiveness is still low, we have to keep our eyes on our IDE hard drive.