RAID technical issues, explain RAID features

RAID-0 (stripe set).

RAID-1 (mirrored volume).

RAID-5 (stripe with distributed parity).

RAID-10 (Mirror Array Stripe).

RAID 0 uses a technique called "striping" to distribute data across disks. There, each "stripe" is spread out into successive "blocks", and the data is divided into chunks ranging from 512 bytes to several megabytes, which are then alternately written to disk. Block 1 is written to disk 1, block 2 is written to disk 2, and so on.

When the system reaches the last disk in the array, it writes to the next segment of disk 1, and so on.

Splitting the data can distribute the load evenly across all drives. Since the drive can be written or read at the same time, the performance is significantly improved. However, it doesn't have data protection capabilities.

If one disk fails, the data is lost. As a result, RAID 0 is not suitable for mission-critical environments, however, it is well suited for the production of images, images, and . Disk usage = 100%.

RAID-1 is also known as a mirror because the data on one disk is completely copied to the other. If there is a data error on one disk, or if there is a bad sector on the disk, then the other disk can remedy the data loss and system interruption caused by the disk failure.

In addition, RAID-1 is duplex – i.e. the entire controller can be replicated, so your data is protected in the event of a disk failure or controller failure. The disadvantage of mirrors and duplex is that they require twice as many drives to replicate data, but the read and write performance of the system does not improve as a result, which can be a significant expense.

RAID-L can be implemented in software or hardware. Disk usage = 50%.

RAID 5 is also known as stripes with distributed parity bits. Each strip has a space the size of a "block" that is used to store parity bits.

RAID 5 distributes parity information across all disks as well. Despite some loss of capacity, RAID 5 is a widely used disk array solution because it provides perfect overall performance.

It is suitable for input/output-dense applications with high read/write ratios, such as transaction processing.

In order to have RAID 5 level redundancy, we need a disk array consisting of at least three disks.

Disk usage ratio = (n-1) n

Introduction to RAID technology, Classification of RAID technology, Development of RAID technology, Specification of RAID technology, Introduction to RAID technology RAID is the abbreviation of Redundant Array of Inexpensive Disks in English, and the Chinese abbreviation is cheap disk redundant array. RAID is a type of hard disk that combines multiple independent hard disks.

Categories: Computer, Networking, >> Hardware.

Problem description: It's better to use plain language that you can understand, otherwise it will make me dizzy.

Analysis: RAID technology will be roughly the state bench cluster to work with multiple hard disks. It's generally more common on servers.

There are seven kinds of RAID, among which RAID0 and RAID1 are the most common, RAID0 works by writing multiple hard disks (if it is two) at the same time, that is, the data stream is divided into two strands, respectively, into two hard disks, which is equal to speeding up (the theoretical speed is equal to twice the read and write speed of a single 7200RPM hard disk), RAID0 is highly respected by some computer enthusiasts, but there is a fatal shortcoming: that is, as long as one of these two hard disks has a problem, will paralyze the system. The parameters of the two hard disks must be exactly the same, that is, they must be the same hard disk.

However, as long as we pay attention to timely backup in daily use, the loss is not big for our ordinary users.

RAID 1 is mostly used on servers, and take those two hard disks as an example, it works by treating one as a pure backup and the other as a normal application hard disk. When writing data, the data system on the ordinary hard disk is automatically backed up to that hard disk, so the advantage is: it improves the disaster recovery ability of the system, and when the main hard disk crashes, the backup hard disk can automatically act as the main hard disk.

Of course, its shortcomings are also obvious: that is, the performance is only equivalent to a single hard disk, and the capacity is only as large as the main hard disk, which is not used by our average users. Others such as RAID 3 and RAID 7 have their own features, but most regular servers don't support them at all, so they don't have much practical value.

In addition, it should be noted that the RAID function needs to be supported by the motherboard, and the motherboard after 05 years generally provides support for the RAID function, that is to say, you only need to buy two hard disks that are the same, and turn on the RAID function in the BIOS (additional software support in the Windows interface is required). If the motherboard doesn't support it, you can only buy a hardware card dedicated to RAID.

Actually, I only know so much, you can search in it, and you will understand.

The function of raid0 is to divide a data into two halves and write them to two hard disks, each of the two hard disks only writes half of the total data, and the time is naturally halved, which is equivalent to the speed increase for users, because the time of writing the same data raid group is half of the theoretical time of a common hard disk of friends, and the game code raid0 does not waste hard disk space, with two hard disks, the sum of storage space is still the sum of two hard disks; RAID1 is equivalent to two hard disks used together, but the data written by each hard disk is exactly the same, the role of RAID1 is real-time hot backup, considering the data security issue, the hard disk storage space of the RAID1 group can only be used half, that is, you buy two hard disks, but the data in the two hard disks is exactly the same, and can only be used as one.

RAID0 With two hard disks, you roughly understand that you have two bottles of 1L water and one bottle of 2L water, assuming that the size of the bottle mouth is the same and the pressure is the same, you check the data as if you take water from it, you pour two bottles together is definitely faster than one bottle. This just explains the speed aspect.

RAID1 is the removal of two hard disks, the things stored in them are exactly the same, and the function of the storage is to prevent a hard disk from suddenly being broken, then you can copy the data from the good hard disk in time, after all, the probability of two hard disks being broken at the same time is very low.

RAID 0 is two hard disks are good, but the storage capacity is the same, and the two hard disks are backed up to each other, which can ensure the security of data!!

Your understanding is basically correct. But there are many more raid levels, just others that are not commonly used.

RAID0 is to combine two or more hard disks into a logical disk, and multiple hard disks are carried out at the same time when reading and writing, so the speed is faster, but because only one of the hard disks fails, the entire array will be finished, and the data security is poor.

RAID1 is to combine an even number of hard disks into a logical disk with a capacity of half of the original capacity, half of the hard disk reads and writes normally, and half of the hard disk backs up data, so the security is high and the speed is not fast.

RAID 5 requires more than 3 hard disks, and the security of data is protected by parity, which can allow one hard disk to fail. Multiple hard disks are involved in reading and writing, so the speed is also good.

It is also worth mentioning that the difference in disk utilization of these array combinations also directly affects their use. Raid0 can use all the disk capacity, Raid1 can use half of it, and Raid5 can use the N-1 number of disk capacity.

As for what you said about raid10, raid50, etc., you are right.

1 All of course is that the more hard disks, the faster the speed. Because doing raid0 is the only way to improve the read speed. However, the advance herein is, first, your array card is PCI-E socket, second, your hard disk should be fast, and third, your motherboard has a free PCI-E socket.

Take PCI array card as an example: PCI is 32-bit, but some motherboards are 33MHz, and the fastest transmission speed of PCI is 133MS in theory, so if you do 2 or 4 RAID0 arrays, the speed is basically unchanged, and it may be a few ms faster. At this time, the flat neck of the speed is on the PCI, and now the array card of the PCI socket will basically be slowly eliminated.

It's just some old motherboards, which are better for server upgrades.

Take PCI-E X1 as an example: The maximum theoretical speed of X1 is 250ms, 2 hard disks as an array, depending on your personal operating environment, the actual speed is generally around 140-160. If you do it with four hard drives, it won't exceed 190.

Generally around 160. The flat neck of the velocity is now on x1.

Take PCI-E X4 as an example: The maximum theoretical speed of X4 is 250M S * 4 = 1000 M S, which is enough to meet the combined speed of the four hard disks. At this time, the speed flat neck is mainly on the hard disk.

So, when you choose an array card, it's best to buy a PCI-E x4 or x8 or x16 socket card.

Like LSI 3041E, Rocket 2640x4, Promise TX4660 these are PCI-E X4's. The most cost-effective is 2640x4

And one more point. Four hard drives are RAID0 arrays, and if your hard drive is, the speed is not all faster than your single-disk SSD hard drive. If your four hard drives are perhaps exceeded.

The above is for reference only and does not represent your actual speed. Thank you for your understanding.

RAID 0 is to separate the data, two hard disks read and write at the same time, the speed is generally several times faster than a few hard disks, but a hard disk is broken, all data is reimbursed.

RAID 1 is used for data backup, the same storage as two hard disks, does not improve performance, but improves security.

There are several modes, there should be a combination of the two, it looks complicated, it is best to find an expert to do it for you.

If I'm not mistaken, the hard drive should be of the same model, otherwise the one with the lowest performance and the smallest capacity will prevail.

Check it out. May help you.

When setting up the RAID function, it seems that you can't choose which RAID groups, and which ones are not grouped. You want to be a group, or you don't have a group.

It is recommended to install a separate SATA RAID card, which is not expensive, and its RAID processing power is definitely better than that that comes with the motherboard. Connect a few blue disk groups to RAID 0.

The motherboard is connected to the SSD as the system.

The IBM server is similar, it's pretty much the same, just follow the prompts.

When the power is turned on to enter the self-test mode, when a prompt appears in the lower left corner to press F1 and ESC, press the ESC key, and the following figure will appear:

Knock Ctrl+H to enter the WebBIOS operation interface, as shown in Figure 1 below(Ctrl+M into pure BIOS).

Click StartEnter the operation interface as shown in the following figure. (If you recognize the HDD on the right).

Click on the Configutation WizardEnter the operation interface with the following options.

Select the second option, New ConfigurationConfigure the new array, as shown in the following figure.

Select the first item, custon configurationConfigure it manually. This is illustrated below.

On the left is a hard disk that can be used as an array. Hold down the Shift key and select the HDD in the Ready state on the left as needed

Click Accept Array to enter the following figure:

Click NextGo to the array configuration page

On the left is the configuration option of the array: raid level: raid 0; stripe size:

64k;rfead policy:read ahead;write policy:write through;cache policy:

cache i/o;select size: (enter the value on the right).Click Accept

Enter Figure 8 below.

Click AcceptAs shown below:

Click yesThe new araay 0 was created successfully. This is illustrated below.

According to the above steps, create a system disk first, select only the 320G hard disk (Figure 6), and the rest of the steps are the same.

Then according to the data disk of 8T in the above step, select all the remaining disks, and the rest of the steps are the same;

One final word: Don't know what your server is for? You must know that RAID0 is only the maximum capacity and the fastest transfer speed, but it is also risky, if one of the 1T disks has a problem, all the data will be lost!!

Suggestion: If 7T disk space is enough, build the data disk into RAID5, RAID5 has a disk space that is dispersed to back up the check data block, when one of the disks has a problem, the RAID card will recover the lost data from the check data block!