Sign in with
Sign up | Sign in
Your question

6 hard drives in raid 0

Tags:
  • NAS / RAID
  • Hard Drives
  • Western Digital
  • Storage
Last response: in Storage
Share
February 27, 2011 9:18:18 PM

is using 500gb 7200 rpm wd drives will 6 of them in raid 0 be faster then 2-3 simple answer please thats all i want.

More about : hard drives raid

a c 187 G Storage
February 27, 2011 10:01:16 PM

Simple answer: no.
-----------------------------------------------------------------------------------------------------------------

Longer answer: it depends.

Random access of short blocks, which is what the OS does 90% of the time will be unaffected, or slower.

If your app uses large sequential file processing, then raid-0 can be faster.

Even then, splitting the inputs from the outputs on separate drives or arrays will do as well, or better.

Synthetic benchmarks will look wonderful, but are irrelevant unless they mimic your workload which is doubtful.


If you want performance, look to SSD drives.

P.S. How do you get your config to show up on the bottom of your post?
m
0
l
February 28, 2011 1:02:03 AM

I put it in my signature in profile options.

I have ssds btw nothing but issues for me i prefer not to use them

but 2 hdds in raid is faster then 1 alone so how is 6 not better then 2 i am just trying to learn here.
m
0
l
Related resources
February 28, 2011 1:04:30 AM

I think i did it wrong though i didnt want raid 0 i wanted it so i had 6 hdds and if i loose one my system wont crash mirror right but how would i set that up with 6 hard drives?
m
0
l
a b G Storage
February 28, 2011 1:54:05 AM

If your looking for data redundency and a performance boost, I'd read up on hybrid raid, more commonly known as raid 10, raid 1+0, etc.

Typically uses four drives, but its capacity is N/2, so 6 1tb drives would give 3tb of space. Adding drives to a Raid 1 or 0 array gives diminishing returns on performance increases due to the increased overhead in determining which drive to read / write too. Raid 1 won't give any improvement to write operations, as each drive will be written to simultaneously. So in reality raid 1 write performance will always be no better than your slowest drive.

Raid 0 will have a speed increase both read and write but won't have too big a jump from 5-6 drives as from 2-3. Again it comes to bandwidth of the controllers and the processing overhead to allocate and organize.

Raid 5 or 6 may be what you want to look at also, but its typically not considered a full redundant array, more single fault tolerant array.

Personally, I'd look into raid 10, or 5 with two drives held in reserve as spares incase a drive fails and the array needs to be rebuilt.
m
0
l
February 28, 2011 3:02:37 AM

Last thing will 4 hdd in raid 0 be faster then 2 for loading games etc?
m
0
l
a c 415 G Storage
February 28, 2011 6:24:30 AM

Since nobody's mentioned it so far I'd like to point out that RAID-0 is not redundant - there is NO protection against disk failure. This means that a four-drive RAID-0 set is FOUR times more likely to fail with complete data loss than a single drive is.

So if you choose to use RAID-0 then either use it only for noncritical data or employ a sound backup strategy to protect it.
m
0
l
a b G Storage
February 28, 2011 6:25:05 PM

cia24 said:
Last thing will 4 hdd in raid 0 be faster then 2 for loading games etc?

Well, yes and no. Throughput will be higher, but so will the processing time to access that data. Also random access time is still just that, and won't be able to outperform a good SSD for raw speed.

In a lab test, 4 drives raid 0 will outperform 2 drives raid 0. Real world, you might get faster by 1-2 seconds, but it won't feel very much faster.
m
0
l
February 28, 2011 6:36:15 PM

someone19 said:
Well, yes and no. Throughput will be higher, but so will the processing time to access that data. Also random access time is still just that, and won't be able to outperform a good SSD for raw speed.

In a lab test, 4 drives raid 0 will outperform 2 drives raid 0. Real world, you might get faster by 1-2 seconds, but it won't feel very much faster.



I could do 5 drives in raid 5, but i hear this will kill write time. But i will have a mirror. I really dont know what to do as long as i back a raid 0 setup with antoher hdd it wont matter if the disk fails. and i dont think that drives just fail all the time i have never had a issue with hdds only ssds.
m
0
l
a c 187 G Storage
February 28, 2011 6:50:05 PM

The value of raid-1 and it's variants like raid-5 is that you can recover from a drive failure quickly. It is for servers that can not tolerate any interruption.

Modern hard drives have a advertised mean time to failure on the order of 500,000+ hours. That is something like 50 years.

With raid-1 you are protecting yourself from specifically a hard drive failure. Not from other failures such as viruses, operator error,
malware, fire, theft, etc.
For that, you need external backup. If you have external backup, you do not need raid-1
m
0
l
a c 415 G Storage
February 28, 2011 8:54:15 PM

> Modern hard drives have a advertised mean time to failure on the order of 500,000+ hours. That is something like 50 years.
...but despite this there are a certain percentage of hard drives (on the order of 3 to 6 percent) which fail prematurely.

> If you have external backup, you do not need raid-1
You don't need it (as much) to protect your data, but if downtime is expensive or very inconvenient then it may still be warranted.
m
0
l
a b G Storage
March 1, 2011 12:12:36 PM

Raid 5 allows you a speed boost, not quite as fast as 0, but allows you to access all of your data after a single drive failure. Raid 5 isn't a mirror, the last drive in the array is used to store parity information. The write time penality is for the system to generate that parity data and store it. So a Raid 5 and a Raid 1 will have the same write performance, possibly slower with raid 5. That data in 5 is stripped across multiple drives however, so you get the read performance of 5 drives.

Raid 0 with six drives increases the likeliness of an unrecoverable loss of data by 6 times. (if any one drive fails, the whole array is lost)

Even with raid 1, 5, 10 etc. A backup solution is needed to ensure valuable data is maintained.
m
0
l
a c 415 G Storage
March 1, 2011 7:42:29 PM

someone19 said:
The write time penality is for the system to generate that parity data and store it. So a Raid 5 and a Raid 1 will have the same write performance, possibly slower with raid 5.
This is absolutely not true.

When you write to a RAID-1 array, the RAID controller issues two write operations, one to each drive, in parallel. Both write operations occur at the same time and you have to wait for whichever one takes the longest.

When you write to a RAID-5 array, the RAID controller has to issue two READ operations - one to the data block about to be overwritten and one to the corresponding parity block. The two reads occur at the same time and you have to wait for whichever one takes the longest. THEN the new parity is computed by XORing the old data block, the parity block, and the new data block, and the RAID controller issues two WRITE operations to update both the data and the parity. The two writes occur at the same time and you have to wait for whichever one takes the longest.

For sequential operations smart RAID controllers can alleviate the extra I/Os to some extent by pre-reading and caching data, but still the basic equation is that RAID-5 writes take AT LEAST TWICE as long as RAID-1 writes.

Do NOT use RAID-5 for write-intensive workloads, it's performance under those conditions really sucks.


> Raid 5 isn't a mirror, the last drive in the array is used to store parity information.

Actually, that's RAID-4, which nobody uses these days because the parity drive becomes a bottleneck for all the other drives. In RAID-5 the parity blocks are distributed across all of the drives to balance the I/O load.
m
0
l
a b G Storage
March 2, 2011 12:12:27 AM

Quote:
> Raid 5 isn't a mirror, the last drive in the array is used to store parity information.

Actually, that's RAID-4, which nobody uses these days because the parity drive becomes a bottleneck for all the other drives. In RAID-5 the parity blocks are distributed across all of the drives to balance the I/O load.


That's what I said shortly after - "That data in 5 is stripped across multiple drives however, so you get the read performance of 5 drives." I realize I'm not exactly completely white-paper technically accurate, but for the purposes of the OP's questions very close to the reality of each implementation.

m
0
l
March 2, 2011 6:28:24 PM

I've decided to put 2 drives in raid 0 and i will back them up daily myself. I uses a jcon program to do this. thanks all for the information.
m
0
l
!