RAID4 and RAID5 Code

RAID4 and RAID5 Code

[SandeepKsinha]

Hi all,

Looks like we have the same code for RAID 4 and RAID 5, I could see
the same make_request for both.

Can someone point me in the code, where exactly we differentiate
between these two. i.e updating checksum block, etc.


-- 
Regards,
Sandeep.





 	
“To learn is to change. Education is a process that changes the learner.”
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: RAID4 and RAID5 Code

[Neil Brown]

On Thursday May 14, sandeepksinha@gmail.com wrote:
> Hi all,
> 
> Looks like we have the same code for RAID 4 and RAID 5, I could see
> the same make_request for both.
> 
> Can someone point me in the code, where exactly we differentiate
> between these two. i.e updating checksum block, etc.

Look for "conf->level" in e.g. raid5_compute_sector.

NeilBrown

Re: RAID4 and RAID5 Code

[H. Peter Anvin]

SandeepKsinha wrote:
> Hi all,
> 
> Looks like we have the same code for RAID 4 and RAID 5, I could see
> the same make_request for both.
> 
> Can someone point me in the code, where exactly we differentiate
> between these two. i.e updating checksum block, etc.
> 

RAID 4 is really nothing but a block layout algorithm for RAID 5 (which
has four more block layout algorithms.)  RAID 6 contains all five under
one "level".

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.

Re: RAID4 and RAID5 Code

[SandeepKsinha]

On Fri, May 15, 2009 at 9:38 PM, H. Peter Anvin <hpa@zytor.com> wrote:
> SandeepKsinha wrote:
>> Hi all,
>>
>> Looks like we have the same code for RAID 4 and RAID 5, I could see
>> the same make_request for both.
>>
>> Can someone point me in the code, where exactly we differentiate
>> between these two. i.e updating checksum block, etc.
>>
>
> RAID 4 is really nothing but a block layout algorithm for RAID 5 (which
> has four more block layout algorithms.)  RAID 6 contains all five under
> one "level".
>
RAID 4 vs RAID 5
Actually debatable. They have their own pros and cons and especially
where you don't consider the physical properties of disk.
Addition of new disk, differences in speed of disks, etc.. create
bottlenecks especially in case of RAID5.

Except hot spot issue with RAID 4, I find it better than all others.
Also, most of the proprietary solutions from IBM, Adaptec and NetApp
suggest RAID4 as compared to others.

No offenses please, this is just an opinion.


>        -hpa
>
> --
> H. Peter Anvin, Intel Open Source Technology Center
> I work for Intel.  I don't speak on their behalf.
>
>



-- 
Regards,
Sandeep.





 	
“To learn is to change. Education is a process that changes the learner.”
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: RAID4 and RAID5 Code

[H. Peter Anvin]

SandeepKsinha wrote:
>>>
>> RAID 4 is really nothing but a block layout algorithm for RAID 5 (which
>> has four more block layout algorithms.)  RAID 6 contains all five under
>> one "level".
>>
> RAID 4 vs RAID 5
> Actually debatable. They have their own pros and cons and especially
> where you don't consider the physical properties of disk.
> Addition of new disk, differences in speed of disks, etc.. create
> bottlenecks especially in case of RAID5.
> 
> Except hot spot issue with RAID 4, I find it better than all others.
> Also, most of the proprietary solutions from IBM, Adaptec and NetApp
> suggest RAID4 as compared to others.
> 
> No offenses please, this is just an opinion.
> 

I was talking referring to how the implementation works.  However, it is
still nothing but a layout policy for the same data, even though it may
have quite different performance attributes.

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.

Re: RAID4 and RAID5 Code

[SandeepKsinha]

On Sat, May 16, 2009 at 12:08 AM, Maurice Hilarius <maurice@harddata.com> wrote:
> SandeepKsinha wrote:
>
> ..
>
> RAID 4 vs RAID 5
> Actually debatable. They have their own pros and cons and especially
> where you don't consider the physical properties of disk.
> Addition of new disk, differences in speed of disks, etc.. create
> bottlenecks especially in case of RAID5.
>
> Except hot spot issue with RAID 4, I find it better than all others.
> Also, most of the proprietary solutions from IBM, Adaptec and NetApp
> suggest RAID4 as compared to others.
>
> No offenses please, this is just an opinion.
>
>
> In a lot of  uses, RAID4 is actually preferable.
> Putting all parity data on one disk allows us to "cheat" and use a faster
> drive for that parity disk.
> Thus we can gain some performance over RAID5 or 6.
>
> To avoid the risk factor increase, we can use a more reliable AND faster
> disk/interface for that parity device.
>
> For reads we gain a fair bit of performance by doing this.
>
> Further, we may "cheat" by using a single larger parity disk.
> By doing so, if we add more data disks, we still have reserve space for
> growth on the larger parity disk.
>
> Yes, our write rates are lousy, but for some uses this is OK.
> Yes, rebuilds are slow too..
>
What makes you think that the rebuilds are slower?

>
>
> --
> With our best regards,
>
> Maurice W. Hilarius         Telephone: 01-780-456-9771
> Hard Data Ltd.                FAX:          01-780-456-9772
> 11060 - 166 Avenue         email:maurice@harddata.com
> Edmonton, AB, Canada      T5X 1Y3



-- 
Regards,
Sandeep.





 	
“To learn is to change. Education is a process that changes the learner.”
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html