Steve Smith, host of your TQA Weekly, talks about Raids 0 through 6, Parity, Mirroring, Striping, Fault Tolerance, protecting your Raid setup, and the differences between hardware and software Raids.
Episode #2-47 released on August 19, 2012
Raid, short for Redundant Array of Independent Drives, is a cluster of hard drives networked together to enhance data security, drive speed, and / or expand the total space of one virtual drive. The virtual drive size and data safety is based on the number of drives and the type of Raid applied to these drives. You must, also, account for the fact that there are two types of raid configurations before setting up the actual Raid type. You have hardware type Raid setups, also the best and safest, and software type Raid setups which lack many of the safety features a dedicated hardware Raid has, like the ability to rebuild a lost hard drive within the Raid easily and in a user friendly fashion.
Let's talk about the 7 most common individual Raid types.
Raid 0 requires 2 hard drives and stripes the data onto two drives, much like putting all the zeros on one drive, and all the ones on the other. This offers speed, but no data security.
Raid 1 requires 2 hard drives has mirroring, it effectively is a safety net for data, but offers no other advantages. The fault tolerance allows for one drive to fail, without losing data.
Raid 2 requires 3 hard drives and has bit level striping, and is capable of recovering one lost drive.
Raid 3 requires 3 hard drives and has byte level striping with dedicated parity, and has a fault tolerance of one hard drive.
Raid 4 requires 3 hard drives and has block level striping with dedicated parity, also, has a fault tolerance of one hard drive.
Raid 5 requires 3 hard drives and has block level striping with distributed parity and has a fault tolerance of one hard drive.
Raid 6 requires 4 hard drives and has block level striping with double distributed parity, and also has a fault tolerance of two hard drives.
What is parity?
Parity is used as error detecting code, the parity bit is the simplest of all the error detecting code.
What is fault tolerance?
The fault tolerance of Raid is defined by the number of hard drives that can fail before the whole of the Raid array fails as a whole. Raid 0 has no fault tolerance, and Raids 1 through 5 only offer a fault tolerance of one hard drive. Raid 6 has a two drive fault tolerance and offers the safest form of Raid to date. Most Raid arrays could fail completely during the rebuild of a new hard drive if another hard drive where to fail. Raid 6 will only fail if two drives containing the same data to create each other fail.
What is striping?
Striping is the logical separation of sequential data, such as files, in such a way that these segments can be accessed across the different hard drives within the Raid array.
What is Mirroring?
Mirroring is the replication of data, exactly as is, on two or more hard drives. What holds true for one drive, must hold true for the other, in other words, a clone.
How does one protect a Raid array?
Now, let's exclude any notion of data security, and encryption, I am referring to how to maintain a safe data environment that prevents any loss of data on any drive, and any environmental damage that could affect a drive like power fluctuations. Using a high quality power supply with multiple rails would prevent any part of the system being under powered, provided you have enough wattage, and it would allow for a constant flow, based on the power input in the power supply. Using an UPS, also known as an uninterruptible power supply, would allow for a constant power flow that would allow the computer or server to safely transmit and store data on any of the hard drives without causing a hard failure, such as a black in the middle of a write procedure. In the event of a hard failure all the drives could fail due to a write failure compromising the Raid array. In the event of hard failure, data retrieval may be impossible. In the case of a single or two drives failing, providing a safe environment for the drives to be recreated is essential for the existence and maintenance of the Raid array. You must always consider using a data backup system for the server such as a second server, tape backup, or cloud storage in the event of hard failure.
Which is better, software or hardware Raid?
Hardware Raid is the easiest to use and normally only requires users to replace the failed drive and replacing it with a new one. Software Raid doesn't always have a way of recreating the lost data, and may compromise the Raid array. In any case, if there is a way to recreate the failed drive in a user friendly fashion, whether by simply changing the drive, or using a recovery tool to aid with the reconstruction, you may use either.
How does one build a Raid?
You will need a minimum of 2 or more identical hard drives, and this is important to note, if you have multiple hard drives with different amounts of storage, the smallest drive becomes the norm, so to avoid any issues, you must always use identical hard drives. Note: using enterprise level hard drives will allow for more data security and longer life of the hard drives. If you using your own personal computer, simply follow the instructions in the manual from there on. Most motherboards can only handle up to Raid 5, if at all. Servers may handle up to Raid 6 or better.
Should I build a Raid for my computer?
What are you trying to do? If it is data security, and you have backups, and are willing to follow the instructions and are able to recover a lost drive, if necessary, yes.
If your trying to have a gain in speed, and are willing to speed hundreds of dollars on drives, and are willing to follow the instructions to recover a lost drive, if necessary, maybe.
If your trying to have a speed increase, and don't understand or have the time to maintain or recover lost hard drives in a Raid, consider using an SSD.
Do I need to use traditional hard drives in a Raid?
No, you can use any hard drive you want, provided you can afford it, and they are all the same size. Using SSD drives require you get rid of virtual memory from writing on them, but they will provide you with tons more speed. If a SSD drive fails, it will cost more however to replace the hard drive.
Next week, I'll be getting into the topic of the sub layer of your web-site, the robot.txt file and site maps, how to create them, how to use them, and what they do.
Remember to like this episode if you were interested in today's topic, share if you think someone else could benefit from the topic, and subscribe if you want to learn more. For the show notes of this episode and others, for more information on other ways to subscribe to our show, to subscribe to our weekly newsletter, and how to participate by submitting your questions, comments, suggestions, and stories, head over to TQAWeekly.com.
Host : Steve Smith | Music : Jonny Lee Hart | Editor : Steve Smith | Producer : Zed Axis Productions