LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
Home Forums Tutorials Articles Register
Go Back   LinuxQuestions.org > Forums > Non-*NIX Forums > General
User Name
Password
General This forum is for non-technical general discussion which can include both Linux and non-Linux topics. Have fun!

Notices


Reply
  Search this Thread
Old 11-25-2005, 11:29 AM   #1
hedpe
Member
 
Registered: Jan 2005
Location: Boston, MA
Distribution: Debian
Posts: 380

Rep: Reputation: 30
SATA150 and Ultra ATA100 performance differences?


Hi,

I was wondering the type of performance differences for SATA150 hard drives and Ultra ATA100 hard drives.

Currently I have two western digital sata150 160gb HD's in a software RAID0. I recently found Seagate Ultra ATA100 200gb HD's on sale for $30 at compusa.

I have ide slots availible and this is a great price. What kind of performance differences could I expect if I put the seagate's in a software RAID0 also?

Thanks!
George
 
Old 11-25-2005, 11:38 AM   #2
Lleb_KCir
Senior Member
 
Registered: Nov 2003
Location: Orlando FL
Distribution: Debian
Posts: 1,765

Rep: Reputation: 45
you can not mix IDE and sATA drives in a RAID, and you are not running RAID you are running FRAID.

the performance difference of IDE vs sATA is about double. the sATA has way higher standards and in a lot of manufactures like Hitachi and Seagate are really just SCSI drives with the ATA controller on them.
 
Old 11-25-2005, 11:43 AM   #3
baldy3105
Member
 
Registered: Jan 2003
Location: Cambridgeshire, UK
Distribution: Mint (Desktop), Debian (Server)
Posts: 891

Rep: Reputation: 184Reputation: 184
For the best info on this sort of stuff go look at www.tomshardware.com
 
Old 11-25-2005, 11:44 AM   #4
hedpe
Member
 
Registered: Jan 2005
Location: Boston, MA
Distribution: Debian
Posts: 380

Original Poster
Rep: Reputation: 30
yes well they would be two completely seperate raids

worst case i can pull them out and set them up and benchmark them
 
Old 11-25-2005, 11:53 AM   #5
Lleb_KCir
Senior Member
 
Registered: Nov 2003
Location: Orlando FL
Distribution: Debian
Posts: 1,765

Rep: Reputation: 45
well that was not what you asked about.

you will see between 5-50% differenance depending on if you are using FRAID or RAID also will depend on if you are using PCI or PCI-X cards and slots for your RAID.
 
Old 11-25-2005, 12:08 PM   #6
hedpe
Member
 
Registered: Jan 2005
Location: Boston, MA
Distribution: Debian
Posts: 380

Original Poster
Rep: Reputation: 30
Quote:
Originally posted by Lleb_KCir
well that was not what you asked about.

you will see between 5-50% differenance depending on if you are using FRAID or RAID also will depend on if you are using PCI or PCI-X cards and slots for your RAID.
actually yes it is

i said I have 2 WD SATA150's in a software raid-0

i said:
"What kind of performance differences could I expect if I put the seagate's in a software RAID0 also?"

the key letter is "A", if i put the seagates in A software raid0 also...

it did not say

"if i put the seagates in the software raid0 also"

they imply two very different things
 
Old 11-26-2005, 08:01 AM   #7
slantoflight
Member
 
Registered: Aug 2005
Distribution: Smoothwall
Posts: 283
Blog Entries: 3

Rep: Reputation: 35
You're insane putting 2 drives like that in a raid0. Give me a drive if you're just going to waste it.
 
Old 11-26-2005, 03:22 PM   #8
Lleb_KCir
Senior Member
 
Registered: Nov 2003
Location: Orlando FL
Distribution: Debian
Posts: 1,765

Rep: Reputation: 45
ok so i read it wrong, but you still have your answer. 5-50% difference depending on if you use FRAID vs RAID
 
Old 11-26-2005, 07:15 PM   #9
slantoflight
Member
 
Registered: Aug 2005
Distribution: Smoothwall
Posts: 283
Blog Entries: 3

Rep: Reputation: 35
Quote:
Originally posted by Lleb_KCir
5-50% difference depending on if you use FRAID vs RAID
Well, thats quite a range is'nt it?
 
Old 11-27-2005, 10:48 AM   #10
Lleb_KCir
Senior Member
 
Registered: Nov 2003
Location: Orlando FL
Distribution: Debian
Posts: 1,765

Rep: Reputation: 45
that is the differance between FRAID and RAID.
 
Old 11-28-2005, 12:31 AM   #11
slantoflight
Member
 
Registered: Aug 2005
Distribution: Smoothwall
Posts: 283
Blog Entries: 3

Rep: Reputation: 35
hmmm.

Could you give us some details maybe?
 
Old 11-29-2005, 01:47 AM   #12
Lleb_KCir
Senior Member
 
Registered: Nov 2003
Location: Orlando FL
Distribution: Debian
Posts: 1,765

Rep: Reputation: 45
http://www.3ware.com/products/serial_ata8000.asp

read up there for starters, hope that is the right place. i got most of my info from one of the big wig sys admins in my LUG. a portion of the info went over my head, but one thing i understand is that with FRAID you are relying on the software and OS to do almost all of the work for your raid.

that means the CPU, RAM, and HD usage plus the work your kernel has to do is greatly increased and also reduces the effectiveness of your raid making it almost worthless, thus only about a 5-10% different if you are using a FRAID vs RAID configuration to get closer to the 30-50% difference.

with a true hardware RAID, keeping in mind that most "hardware raid cards" are not true hardware raid, but are in reality FRAID cards, you will get a dramatic performance difference as you can add extra RAM to the real RAID card and 100% of the processing is done on the card and NOT via the OS, or other parts of the computer to slow things down.

that will put the bottle neck on your bus (PCI-X bus). that will be more then high enough to deal with most traffic even in a high demand LAN.

here is a portion of his explanation for the RAID out of an e-mail:

Quote:
98% of the hardware cards are _not_ hardware RAID. They are what I call FRAID -- fake/free RAID. They do 100% of RAID in software, and they _suck_.

For Linux, you want 3Ware -- period. Pay the $100, 250 or 400,
depending on if you want 2, 4 or 8 SATA channels. If you only need
400-800GB on your file server, go for four (4) 200-400GB drives on a
3Ware Escalade 8506-4LP card in RAID-10. They are _very_worth_it_!

And I'd pay the $100 for a 3Ware Escalade 8006-2 with two (2) redundant SATA in RAID-1 drives on any other systems. Trust me on this, you don't want to go down and not know about it. 3Ware has the ultimate Linux compatibility and monitoring features.
here is a bit about sATA vs SCSI

Quote:
>Why sATA vs SCSI for a file server?


It's not SATA v. SCSI. You're still using an _intelligent_ RAID controller, that's the big deal. Furthermore, if you're really stuck on SCSI, you can get the _exact_same_ Hitachi 10K SCSI models as Western Digital Raptor 10K SATA drives.

SCSI is starting to be replaced by Serial Attached SCSI (SAS). SAS is
basically SATA with the SCSI-2 command set atop. SAS can use either
SATA or SAS drives.


>> I did only a little read up on RAID-10 (first time i have heard of it)
>> and it still looks like it will only give standard single read/write
>> vs SCSI with multiple read/write.


RAID-10 is _faster_ and _higher_performing_ than RAID-5 -- _regardless_ of ATA, SATA, SCSI, SAS, etc... Unless you need maximum storage, RAID-10 is _always_ better than RAID-5 -- especially for multiple reads/writes.

3Ware Escalade ASICs queue up more I/O than 98% of the SCSI RAID
controllers out there. Unless you want to drop $700+ on an Scale-based U320 SCSI card, a $300-400 3Ware Escalade is a heck of a lot better. Heck, a $600 XScale-based SATA card is typically better.

>Am I missing something here? yes sATA drives vs SCSI are much less expensive when it comes to storage space.


No, don't think of it as SATA v. SCSI. That is a farce. People don't
realize that some SATA and SCSI models are the _exact_same_ drive these days.

Think of it of as "enterprise" class storage (18, 36, 73, 146GB) drives
versus "commodity" class storage (200, 300, 400GB, etc...) drives. You can get _both_ in SATA or SCSI.

Now there's also "24x7 commodity" drives in the Seagate NL35, Western Digital RE, etc... They are commodity capacity, with 24x7 operation/5 year warranty.

Again, even if you go SATA, you can _still_ get 73GB 10K drives with
Western Digital Raptors. They are the _exact_same_ drives as the 73GB 10K Hitachi UltraStar 10K.

so with your sATA drive in a proper RAID you are running just as good as if you were in a SCSI, and better under most cases, so to try to get IDE to compete with that is just unreasonable.

if you are running a FRAID (software RAID) then you will not see a noticeable difference thus the 5-10% range.

hope that is semi clear.
 
  


Reply



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
RHEL driver for controller PROMISE SATA150 TX2Plus mngaldi Linux - Hardware 2 06-22-2006 12:44 PM
Not another SATA150 TX4 thread cyberpuerka Linux - Hardware 0 05-06-2005 03:32 PM
ultra-2 and ultra-3 SCSI drives on the same bus? PsychosisNode Linux - Hardware 3 12-25-2004 12:52 AM
config Promise SATA150 TX2 plus on FC2 wigdog2 Fedora 4 09-19-2004 03:02 AM
ata100.i - self compiled kernel doesn't recognize ATA100 controller... wtf? keex Slackware 1 06-28-2002 03:57 AM

LinuxQuestions.org > Forums > Non-*NIX Forums > General

All times are GMT -5. The time now is 09:12 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration