Subtitles of the Movie

Exam objective 2.204.1 has a weight of 2, and verifies that candidates are able to configure and implement software . This includes using and configuring s 0, 1 and 5. The word is an acronym. The letters stand for Redundant Array of Inexpensive Disks. Of course, the primary reason for adding disk drives is to add disk space, there are two reasons for adding those drives as drives. If the devices are configured properly, it will make reading and writing data faster. Also, it is possible to configure the disks so they have a built in redundancy. Everything written goes to two or more disk drives. Which of these advantages you get, and just how you get them, is determined by the way you configure your . The most rudimentary configuration is linear mode. The disks act as one large file system. The disks are seen as a chain. As soon as one disk is full, data is written to the next one. The scheme has no redundancy, the loss of a disk drive will corrupt the data. With O, the data is distributed across the disks. When you write a file to disk, a piece of it is written to each drive, and it's distributed across the platters of each drive in such a way that it minimizes head movement, and maximizes speed. That type of storage on multiple platters of a disk drive is known as disk striping. Speed is the prime advantage of 0. It has no redundancy, but without redundancy, you get the full advantage of the space of all drives. For example, if you have four drives, 15 gigabytes each, the total storage space is 60 gigabytes. With 1, the same data is written to more than one disk. This type of duplication is also known as mirroring. If one drive in the array fails, the other drives still contain data. If the bad drive is then replaced, it will catch up by being filled with data from the other drives, so no data is lost. It's not as fast as O, because the controller has to write the data to multiple drives. It also is smaller than O. The amount of storage space is no larger than the smallest disk drive in the array. That's because every drive duplicates all data. 4 is an attempt to combine the advantages of 0 and 1. It has one disk dedicated to error correcting, or parroting information. It requires a minimum of three disk drives: two for data, and one for the parity information. This way it provides redundancy. It can stand a loss of only one drive. It uses the content of the other two to recreate the lost information. The result is some of the speed of 0 with some of the redundancy of 1. All of the data is striped, except the parity drive, and that's the reason 4 is not used as much as 5. 5 is the same as 4, except the data on the parity disk is striped for speed, so this one is as fast as 0 with the redundancy of 4. Like 4, it can only survive the failure of one disk drive. 6 is very much like 5, except that it can lose two drives without losing any data. It has two parity drives and uses striping for those parity drives, so it's fast also. A Node is configured by having an entry in the table file. This example entry is configured as a level 5. It has three disks and one spare. If a disk fails the spare will be used in its place. The disk in this array have a Super Block, which is used to keep things straight and prevent the disks from getting confused. Level 5 has a parity disk and this entry specifies the type of parity to be used. The rest of the entry lists the actual device nodes of the devices themselves for the disks that make up the three primary drives and the spare. There is a very flexible software tool you can use to work with MDADM is for managing MD devices: devices. It can be used to construct disk arrays, change the sizes of arrays, monitor the condition of arrays, and other things. This file contains the configuration information needed for managing the disk arrays. If you have a system with multiple disk drives that you can use to experiment, you should be able to set them up as a system. If not, you should locate the MAN pages for MD and MDADM and read those.