[K12OSN] Linux "Software RAID"
David Hopkins
dahopkins429 at gmail.com
Sat Aug 9 13:19:28 UTC 2008
I"ve enjoyed reading this thread, but have a question: I thought one
of the reasons to continue using RAID 5 vs RAID 1 was speed. Tests on
my systems show that the RAID 5 I/O is at least 3x the speed of RAID
1. I have hardware based RAID (either 3ware SATA controllers (large
disks) or Adaptec SCSI). I put the OS on a RAID 1, and all data (e.g.
/home) is on a RAID 5. I know I could get very cheap SATA drives and
my controllers support mirroring so perhaps that is the next step.
Sincerely,
Dave Hopkins
On Fri, Aug 8, 2008 at 3:24 PM, Les Mikesell <lesmikesell at gmail.com> wrote:
> Rob Owens wrote:
>>
>>> raid5: automatically using best checksumming function: pIII_sse
>>> pIII_sse : 2604.000 MB/sec
>>> raid5: using function: pIII_sse (2604.000 MB/sec)
>
>> I don't see any of that info on my systems with software RAID 1.
>
> Raid1 doesn't do any calculations - it just tells the hardware to write
> twice and reads from whichever copy is faster. Any performance hit comes
> from CPU involvement during writes (not much unless you are on old IDE or
> USB connections) or overloading the PCI bus.
>
>> What
>>
>> do those values mean? Are they read speads? Write speeds?
>
> That's the math to compute the parity that needs to be written along with
> the data. And if you have a failed drive in the set it will do a similar
> computation to reconstruct the data in degraded mode (where raid1 just reads
> the other copy at full speed).
>
> --
> Les Mikesell
> lesmikesell at gmail.com
>
> _______________________________________________
> K12OSN mailing list
> K12OSN at redhat.com
> https://www.redhat.com/mailman/listinfo/k12osn
> For more info see <http://www.k12os.org>
>
More information about the K12OSN
mailing list