UnRaid vs Raid

[Shehryar23]

Can someone please explain me what is UNRAID. If Linus has a video on what is unraid can you share the link if not can you please explain it. Can you also tell what is the difference between raid and unraid. i am a newbie in this

[ARikozuM]

LimeTech UnRAID is an operating system. It allows you to set up a NAS.

 

RAID stands for redundant array of independent disks. 

 

These aren't two types of RAID. 

[Kodiak.]

UnRAID is an operating system in regards to virtualization and Network Attached Storage.

 

RAID stands for "Redundant Array of Independent Disks." It functions by putting drives together in a "singular" array.

[Drak3]

UnRAID is an operating system designed to run a NAS and a second OS via a type of virtual machine called KVM.

 

Basically, if you had a hexacore CPU, ypu could run Windows off of 4 as if it were its own computer, and the other two could be a NAS back up.

 

Or, you could just run multiple VMs. If you had a Decacore, you would do 4 cores for virtual machine A, 4 for virtual machine B, and 2 for UnRAID.

[NelizMastr]

RAID will allow you to put a number of disks in an array, that means you can set them up to work together. You can optimize this array for either performance or reliability with different RAID levels. RAID0 will combine all drives for max performance, where RAID6 will use extra disks for data protection, reducing capacity.

 

UnRAID is just an operating system meant for use as a network and virtualization appliance.

[Lady Fitzgerald]

The purpose of RAIDs (other than RAID 0) and unRAID is to provide redundancy. Redundancy only protects against data loss due to drive failure (up to a point), which will allow a computer to continue to operate after a drive failure, but should NOT be considered a backup (in fact both RAID and unRAID need to be backed up to protect from causes of data loss not due to drive failure). RAID and unRAID have overhead, meaning the total usable drive capacity will be less than the combined capacity of all the drives. There a many levels of RAID but the most common in use for redundancy are RAID 1, 5, and 6 (RAID 0 is only for increasing speed and combining drives into a single volume). Nested RAIDs, such as 10, 50, and 60,  combine the advantages of more than one kind of RAID (usually RAID 0 and another RAID level) to make up for slower speeds and to improve drive failure tolerance (at a cost of higher drive overhead).

 

RAID levels (other than 0) store data on multiple disks in stripes with the stripes distributed in such a way that, if one or more disks (the number depending on the level of RAID) should fail, all the data will still exist distributed across the remaining drives. The downside of RAIDs is, if more disks fail than the RAID is intended to allow, all the data will be lost. Also, drives need to be all the same size. If drives of various sizes are used, every drive will only use the capacity of the smallest drive. Adding drives to a RAID involve rebuilding the entire RAID

 

UnRAID doesn't use striping. Instead of being scattered across stripes, data is written directly to the drives and one or two parity drives provide the redundancy. To over simplify the process, when using a single parity drive, the unRAID system will start by looking at the first bit of all the nonparity drives and add up the 0s and 1s. If the total is even, it will write a 0 in the first bit of the parity drive. If the total is odd, it will write a 1. Rinse and repeat until all bits of data have been accounted for. When a data drive fails, unRAID will start recovery on the replacement drive by again adding up the first bit of each drive, then compare it to the first bit of parity drive, then write either a 0 or a 1 to the replacement drive to make the total match the parity drive again. Again, rinse and repeat for the remaining bits of data.

 

UnRAIDs that use two parity drives use a complex algorithm to accomplish the same thing as a system with one parity drive. The advantage of two parity drives is, if one parity drive should die, the unRAID will still be protected. Losing the pairity drive on a one pairity unRAID will not cause data loss but the data will be unprotected until the pairity drive can be replaced and rebuilt.

 

The advantages of unRAID include being able to use drives of different sizes as long as the parity drive(s) is as large as or larger that the largest data drive. Since data is written directly to each drive instead of scattered across multiple drives, if more than the allowable number of drives fail, the remaining drives will still have the data written on them fully accessible, minimizing data loss. Unlike RAID, adding a drive does not require rebuilding the entire array.

 

The main downside of unRAID is it's painfully slow, often too slow to keep up with the incoming data. One way to deal with that is to use SSDS for all the drives. That can be rather expensive. A less expensive way is to use a cache to temporarily hold incoming data until the unRAID catches up. While HDDs can be used for that, SSDs are usually a better choice. This works only where there are sufficent lulls in the incoming stream of data to allow the unRAID time to catch up.

[Shehryar23]

On 10/8/2018 at 7:56 PM, Lady Fitzgerald said:

Since data is written directly to each drive instead of scattered across multiple drives, if more than the allowable number of drives fail, the remaining drives will still have the data written on them fully accessible, minimizing data loss

Wont the parity drive be used to rebuild all of the data that was lost in multiple drives. Or parity drive can be only used to rebuild data if one drives fail not more then one drive fail. meaning if 1 drive fail we can use parity drive to rebuild that drive data, if two drive or more fail can we still use parity drive to rebuild all the data from those two or more drives

[Lady Fitzgerald]

1 hour ago, Shehryar23 said:

Wont the parity drive be used to rebuild all of the data that was lost in multiple drives. Or parity drive can be only used to rebuild data if one drives fail not more then one drive fail. meaning if 1 drive fail we can use parity drive to rebuild that drive data, if two drive or more fail can we still use parity drive to rebuild all the data from those two or more drives

In an unRAID with one parity drive, only one drive at a time can fail and still be recovered from the pairity drive. If two drives fail, they cannot be rebuilt using the parity drive.

 

If an unRAID has two parity drives, up to two drives can fail and still be rebuilt from the parity drives.

 

Keep in mind that parity drives do not actually store your data. They store information that can be compared to the remaining data in the array to determine what was on the failed drive(s).

 

Unlike a RAID, where data is fragmented across multiple stripes on drives across the array, data in unRAID is writen directly to a single drive until it's full then moves to the next drive unless all drives are approaching full and become fragmented. It's also possible to allocate specific data to be written to specific drives. Since data isn't deliberately scattered across multiple drives, one usually can just remove a drive from the array and still access the data there as though it was never in an array. The advantage there is if the number of drives that failed exceed what the unRAID can protect, only the failed drives would need to be subjected to expensive, professional data recovery, unlike RAID, where all drives would need be subjected to expensive, professional data recovery.

 

While RAID and unRAID will protect against drive failure, up to a point, there are ways data can be lost, such as viruses and other malware, theft, power surges, accidental deletion, etc. The only reasonably reliable way to avoid data loss for whatever reason is to backup the unRAID or RAID.

[Shehryar23]

8 hours ago, Lady Fitzgerald said:

In an unRAID with one parity drive, only one drive at a time can fail and still be recovered from the pairity drive. If two drives fail, they cannot be rebuilt using the parity drive. 

So basically it same thing as RAID if one drive fails parity from the other drive will be used to rebuild the data but if more drives fail then all data is gone. UNraid is same one drive gone use parity drive to rebuild data 2 drives gone (we can rebuild one drive from parity and the other drive data is gone am i right??)

and can you also share few links where i can learn more about raid and unraid i am creating my plex server and the thing i am stuck is at the OS what to use so i need to understand unraid first it would be great if you can help

[Lady Fitzgerald]

1 hour ago, Shehryar23 said:

So basically it same thing as RAID if one drive fails parity from the other drive will be used to rebuild the data but if more drives fail then all data is gone. UNraid is same one drive gone use parity drive to rebuild data 2 drives gone (we can rebuild one drive from parity and the other drive data is gone am i right??)

and can you also share few links where i can learn more about raid and unraid i am creating my plex server and the thing i am stuck is at the OS what to use so i need to understand unraid first it would be great if you can help

Someone else is going to have to dive in here because it appears I've been unable to explain this to you adequately.

[Salacious B Crum]

unRAID's other main advantage besides ability to use different size disks in the array is that you only allocate 1 or 2 parity disks and all other drive space is available for data storage unlike RAID where depending on your choice i.e. RAID 1, 5, 6, 10 allot more space is allocated to the configuration for an example RAID 1 requires a minimum of 2 disks say 4TB each but your data storage capacity will only be 4TB as your data is mirrored across both disks also its quite easy to just toss another disk/s into an unRAID array