Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

RAID 4 in Netapp

vas_3
Advisor

RAID 4 in Netapp

HI All,

I hve one query regarding RAID4.

1.Why Netapp using RAID 4?.RAID 4 has any advantages compated to RAID5?.

Regards
Sri.
8 REPLIES
VK2COT
Honored Contributor

Re: RAID 4 in Netapp

Hello,

a) RAID-4 uses large stripes. It can read
records from any single drive. This allows
you to take advantage of overlapped I/O for
read operations. Since all write operations
have to update the parity drive, no I/O
overlapping is possible.

b) RAID-5 includes a rotating parity array,
thus addressing the write limitation in
RAID-4. All read and write operations can be
overlapped. RAID-5 stores parity information
but not redundant data (however, parity
information can be used to reconstruct the
data). RAID-5 requires at least three and
usually five disks for the array. It's best
for multi-user systems in which performance
is not critical or which do few write
operations.

c) RAID-4 is faster than RAID-5 because the
parity is written to a dedicated disk,
rather than scattered around with the data.
If you have N disks, you'd have N-1 data disks, and 1 parity disk.

What happens when the parity disk is lost?
You have lost all protection! There is no
parity, and hence no way to calculate
the data.

For this reason RAID-4 is considered
dangerous and is not often used.

NetApp uses it in their WAFL file system for
their product line of NetApp Filer and
NetCache.

d) So, the advantage of RAID-4 is that it
has the highest read data transaction rate,
with a medium write data transaction rate.

Data is stripped on disk creating high
efficiency along with a good aggregate
transfer rate. Parity is stored on a
separate disk.

RAID-4 offers no other advantages over
RAID-5.

Cheers,

VK2COT
VK2COT - Dusan Baljevic
TTr
Honored Contributor

Re: RAID 4 in Netapp

A couple of things.
First, VK2COT, with all due respect you are not making any sense in you reply above. RAID4 is not faster than RAID5 and it is not dangerous. Take a look at
http://en.wikipedia.org/wiki/Raid4#RAID_4
for an explanation of RAID4 and RAID5. As you can see they are identical in terms of data and parity disks. The ONLY difference is the parity disk in RAID4 is dedicated and the parity disk in RAID5 is distributed. And that's all the difference there is.

Now the distributed parity disk in RAID5 gives an advantage to RAID5 over RAID4 for the following reason.

Lets say you have a 5-disk raid4 and a 5-disk raid5 group.

In a read operation, the parity disk does NOT needed to be read. So reading a raid4 set involves always reading the first 4 disks, 1,2,3,4 and never disk 5.
Reading a raid5 disk still requires reading 4 disks, BUT a different set of 4 disks each time. One read would read disks 1,2,3,4 another read would read disks 1,2,3,5, another one 1,2,4,5 and so on. This is the BIG advantage of RAID5 over RAID4, you are doing read i/o using 5 disks in RAID5 and 4 disks in RAID4. In the example above with 5 disks, raid5 would outperform raid4 by about 20% (the equivalent of the fifth disk)

A similar argument can be made for writing. In both cases all 5 disks are written to but in raid4 the 5th disk is always calculated as a parity disk and the writes are queued differently for the data disk and differently for the parity disk. In a write intense raid group you may end up with a bottleneck on the parity disk as the parity stripe is calculated to be written out. In the case of raid5 since the parity disk is distributed, the writing is much more uniform across all 5 disks.

Historically raid4 came first and it was a huge innovation. Then when raid5 came out as an improvement over raid4, it phased out raid4. And that's that.

Now back to NetApp, I don't thing they use RAID4, instead they use RAID6, two parity disks and they do that for extra safety. There are two kinds of RAID6 depending on how they treat the two parity disks and one raid6 type is much faster than the other. But that's another thread topic.
IBaltay
Honored Contributor

Re: RAID 4 in Netapp

Hi,
this could help you:
http://www.redbooks.ibm.com/redpapers/pdfs/redp4169.pdf

the pain is one part of the reality
IBaltay
Honored Contributor

Re: RAID 4 in Netapp

Hi,
as well as chapter 5. of the following doc:http://www.redbooks.ibm.com/redpieces/pdfs/sg247129.pdf
the pain is one part of the reality
VK2COT
Honored Contributor

Re: RAID 4 in Netapp

Hello,

I still stand by statement that RAID-4 is DANGEROUS because the parity is NOT DISTRIBUTED and IS ON SINGLE DISK.
Wikipedia image clearly shows it :)

Loose the parity disk and everything is
"unconnected".

NetApp is known to support RAID-0, RAID-1+0,
RAID-4, and now RAID-6 (RAID-DP). The RAID-6
by NetApp came in 2006 after HDS and HP
offerings in their product range.

Cheers,

VK2COT
VK2COT - Dusan Baljevic
Uwe Zessin
Honored Contributor

Re: RAID 4 in Netapp

I don't see how distributed parity gives you an advantage if you have two disk drives fail - no matter if its RAID-4 or RAID-5.

The advantage of RAID-4 in a NetApp box is:

- the file system and the RAID-level work together - is is not a file system on a RAID block device.

- an aggregate (RAID-set) can be grown easily by adding a 'zeroed disk'. The file system just starts using the additional disk drive. Recovery is no problem because Any-Data XOR 0 = Any-Data.
.
TTr
Honored Contributor

Re: RAID 4 in Netapp

> I still stand by statement that RAID-4 is DANGEROUS because the parity is NOT DISTRIBUTED and IS ON SINGLE DISK.

Actually if THE parity disk fails in raid4, there is NO need for data reconstruction while the disk has NOT been replaced yet, whereas in RAID5 you need to do data reconstruction for most of the stripes. So raid4 has an advantage over raid4 in this case.
As far as safety, raid4 and raid5 can tolerate a disk failure the same way. It is the floating parity in raid5 that gives raid5 a performance advantage over raid4. If you read carefully mu explanation above you would understand it.

> Loose the parity disk and everything is
"unconnected".

What does this mean? Do you read or understand what you write? Losing the parity disk in raid4 does not disconnect anything differently tahn losing a data disk in raid4 or any disk in raid5.
gregersenj
Honored Contributor

Re: RAID 4 in Netapp

http://www.baarf.com/

Here's you will find a link to the original RAID study at Berkely Uni.

Cheers
/jag