[K12OSN] SATA drive for server with 25 users?

[James P. Kinney III]

On Mon, 2006-10-16 at 19:46 -0700, Robert Arkiletian wrote:
> 
> 
> On 10/16/06, Petre Scheie <petre at maltzen.net> wrote:
>         We haven't had any discussion for a while now as to how well
>         SATA drives scale up in a
>         K12LTSP server.  It used to be, back in the PATA days, that an
>         ATA (also known as IDE)
>         drive would handle up to 10 clients, but going any higher than
>         that resulted in poor 
>         performance that could be addressed only by going to SCSI with
>         its ability to re-order
>         queues and so forth.  But SATA has been out for a while, it
>         now has many of the features
>         of SCSI, and I see that 10K RPM versions are available, and so
>         I'm wondering if the 
>         consensus now is that SATA is good enough for small and even
>         mid-size servers, where by
>         'midsize' I mean roughly 25 clients hanging off of it.  What
>         about 7200RPM SATA drives?
>         Thoughts?
> 
> Although I use scsi in my server my advice would be to make a software
> raid 1 for / , /var, /tmp on 2 fast 10k rpm drives and /home and swap
> on another 2 (bigger) 7200rpm drives. Intel 945P chipsets usually come
> with ICHR southbridges (which have AHCI needed for NCQ) that have 4
> SATA ports. Or if you want lots of space for /home then buy another
> cheap Sil3132 based PCI-E SATA controller (I think SYBA makes them)
> for another 2 SATA ports and make a RAID 5 for /home with upto 4
> drives. 
>  
Something to keep in the back of your head: RAID 1 is great. It has
excellent speed on reads as it can send the heads in different
directions and part of the file from drive A and another part from drive
B.

NFS can't handle a drive hiccup error during a read from a mirror drive
that is failing. It will cause a hard lock of the clients. I'm not done
on the diagnostics yet to determine if the kernel passed a drive read
error through the RAID process and into the NFS delivery process but it
was a disaster no matter how it happened. I'm still considering it as a
kernel fault (software RAID) as it should have trapped the error, failed
the drive and then reverted to the other drive.

> 
>         Petre
>         
>         _______________________________________________
>         K12OSN mailing list 
>         K12OSN at redhat.com
>         https://www.redhat.com/mailman/listinfo/k12osn
>         For more info see <http://www.k12os.org>
> 
> 
> 
> -- 
> Robert Arkiletian
> Eric Hamber Secondary, Vancouver, Canada
> Fl_TeacherTool http://www3.telus.net/public/robark/Fl_TeacherTool/
> C++ GUI tutorial http://www3.telus.net/public/robark/ 
> _______________________________________________
> K12OSN mailing list
> K12OSN at redhat.com
> https://www.redhat.com/mailman/listinfo/k12osn
> For more info see <http://www.k12os.org>
-- 
James P. Kinney III          
CEO & Director of Engineering 
Local Net Solutions,LLC        
770-493-8244                    
http://www.localnetsolutions.com

GPG ID: 829C6CA7 James P. Kinney III (M.S. Physics)
<jkinney at localnetsolutions.com>
Fingerprint = 3C9E 6366 54FC A3FE BA4D 0659 6190 ADC3 829C 6CA7
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/k12osn/attachments/20061017/56fa24a1/attachment.sig>