Relative hard drive configuration performance

An occasional client is experiencing a server failure and their improvement in the performance of their LOB had me wondering relative performance based on drive configuration.


Their situation is the LOB vendor supplied the server, a Windows Server 2003 computer, and was supposed to be monitoring it. They have 4 hard drives in it. 3 make up a RAID 5 array and the 4th was supposed to be a hot spare. Recently we noticed that 1 of the 3 drives failed and the 4th did not kick in. Turns out they hadn't configured the RAID controller to email them when there was a problem and the 4th drive wasn't set up to be a hot spare, We set up the hot spare but now it seems another drive is failing, while we wait for the replacement drive. And the server has become unstable, rebooting during the day.


They moved the LOB to a desktop computer while they send replacement parts. Good thing it's a small office as the Win Vista Business only allows 10 connections.


The server is still up most of the time and is doing DHCP and DNS. To the point of the question, they say performance of the LOB has improved.


In general and speaking in relative terms, would anyone care to comment on relative performance of different drive configurations?


a single drive

Hardware RAD 1

Hardware RAID 5

Degraded Hardware RAiD 1

Degraded Hardware RAID 5

Hardware RAID 6

Hardware RAID 10


Certainly this specific situation has many other variables since its an entirely different OS, processor, RAM, et al. I am just curious how much improvement could be attributed from going from degraded RAID 5 to a single drive. And yes, the quality of the hardware likely can have an influence. I am speaking in general terms though.


Thank you
BeGentleWithMe-INeedHelpAsked:
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

DavidPresidentCommented:
Unfortunately, what you are asking for will provide absolutely meaningless results unless you add additional constraints.    The type of I/O (random vs sequential vs large block vs small block vs reads vs writes vs cached vs non-cached vs saturated I/O channel vs non-saturated I/O channel) matters.  SAS vs SATA matters. Number of disks matter.  Firmware and architecture of controllers matter.

You might as well ask if a Mac vs a PC will run your software faster.   If you want something you can use, provide a specific configuration.  Otherwise, the answer is "it depends".
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
DavidPresidentCommented:
In fact, even a degraded RAID array will outperform a non-degraded RAID array in write intensive applications.  (Why? No RAID write overhead to worry about).    

But this changes if you have a raid controller with a battery back up that disables the write cache if it is degraded.  (On some controllers you can override this, so now we even have to look at configurable settings for the specific controller.  So I can't even generalize this answer.

P.S. filesystem settings matter too.  How many blocks of data does the file system read/write at a time, and compare that to how the RAID controller reads from the disk.  Does the disk read/write 64KB at a time, but your O/S reads/writes 4KB at a time?  If that is the case then no matter what, you're doing 8X the I/O on the disk then is necessary and just throwing performance away.
0
BeGentleWithMe-INeedHelpAuthor Commented:
OK, again, in general

Would you be surprised  moving an LOB from server with server 2003 and degraded raid 5 to win vista 32 bit business low / medium end desktop with single drive would get noticeable  better performance?

or with those admittedly vague specs, would the answer be "got a coin"?
0
Powerful Yet Easy-to-Use Network Monitoring

Identify excessive bandwidth utilization or unexpected application traffic with SolarWinds Bandwidth Analyzer Pack.

BeGentleWithMe-INeedHelpAuthor Commented:
The server has more ram than the desktop, it's running server OS. Bueller?
0
DavidPresidentCommented:
Read performance on the degraded RAID5 will be faster with the degraded RAID5 assuming same make/model of disks.  Write performance on a single disk probably a little faster.

But since this is also the boot disk with swap and log files, then it will be doing more writes.

So, gotta coin is the right answer.  Sorry.

But do this instead, destroy the RAID5 and go to a 2-disk RAID1. That will be faster real-world in most likely I/O situation for a W2K based server.
0
BeGentleWithMe-INeedHelpAuthor Commented:
thanks
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Microsoft Server OS

From novice to tech pro — start learning today.