Raid 5 Array fails to rebuild with new disc installed

Smithds21 used Ask the Experts™
Let me start by saying that I have no detailed experience of RAID 5 arrays.

We have a 3 Disc, single volume, RAID 5 array operating an SBS 2008 Server.  The RAID controller is Intel Matrix Manager accessed via Intel Matrix 8.9, Storage Controller.

The Port 2 disc was reported as failed and subsequently I replaced it.  However, although the new disc was a 500gb Seagate Barracuda, as per Ports 0 & 1, it was a newer version as the original ST3500418AS are no longer available.

The RAID controller recognised the new disc and I took the option to include it in the volume.  However, it was only after the 'rebuild' had started that I noticed that the physical sector size of the new disc was 3024 bytes and not 512 bytes per the other 2 discs.  I let the 'rebuild' process run to completion and the new disk was added into the array at Port 2.  

All appeared well until I tried to reboot the server, it would not boot, simply cycling round and round past the Intel boot screen and appearing to start to load but after 2 minutes or so it rebooted.

I refitted the original failed disc and the file server rebooted, per normal, but of course with a degraded array.

I contacted the tech support desk of the well known UK retailer of the disk and was advised to reformat the disk on a separate PC to 512bytes and then re-introduce it to the RAID 5 array.  

I did as advised but the file server would not boot with the new disc simply connected to the array port, never mind being introduced to the array, I never got that far.  I am now again running in a degraded state with the old disc in place whilst trying to find a solution.

The 'techie' further advised that my only option would be to rebuild the array, but this is with the inferred loss of data and I do not really want to do that.

Any advice would be most welcome.  I attach screen dumps of the array as currently seen via the Matrix Storage Controller.
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Sector size is imbedded in the disk controller mounted on the hard disk. New Seagate drives are available with 4096 bytes/sector and 512 bytes/sector. You need a 500GB disk set for 512 bytes/sector. Only option I can see is to replace the disk with one that is compatible with the other 2 disks.
Those drives were NEVER meant to be used in a hardware RAID 5 scenario.
They are desktop drives and have recovery mechanisms meant for desktops,not servers.
All of those drives should be enterprise class (ES) if you value your data.

A short explanation ,desktop drives ,if they hit a bad spot on the disk,can take up to 30 seconds to respond the controller because they are attempting  recover a bad sector.
RAID enabled disks will time out after 5 to 7 seconds and if the controller get no response,it will flag the drive bad.
Picture two desktops drives both taking up to 30 seconds to respond,no more RAID and your data is gone.

Same thing in a rebuild scenario.
Angular Fundamentals

Learn the fundamentals of Angular 2, a JavaScript framework for developing dynamic single page applications.


In response to pgm554.

Our previous server ran SBS 2003 on a RAID 0 array and in almost 10 years of service it never skipped a beat.  Since day one of commissioning the new server running on SBS 2008 with RAID 5, built by a well known UK specialist retailer, we have had to endure very slow system performance.  So bad in fact that we had the retailer send techies to help find the problem, they never did and so we have always been in a retrograde condition since day 1.

Coming specifically to your advice about ES class disks is there methodology that you could suggest that would allow me to ghost the current discs onto replacement discs, alleviating the need to formally rebuild the server software, the apps software and the data?  A big ask I guess!  

We take daily tape full back ups which contain sql databases, ms exchange server data and
operational/business systems data and I guess the last resort is to rebuild from scratch.  Don't really want to do this if at all possible.  If I have go to that trouble I think I would prefer to revert to RAID 0 as in our SBS 2003 days.
That Intel "RAID"  controller is nothing more than a glorified piece of software built on board to the MOBO.

0 is a far different beast than 5.

Parity stripes are written across all disks and the controller makes the decisions as to how the array is controlled and how the disks are considered good or bad.

If you do not use the right disks and controllers,you will not be a happy camper

RAID 1 is a lesser of 2 evils if you are going to use desktop drives.

There are products out there that allow for images of the system to be cloned or copied.

Appassure,Storagecraft,Paragon ,and Symantec all have imaging based products that will allow you to image or migrate your old array to a new one without having to install everything from scratch.

It might behoove you to start from scratch because of the SQL database that you are running.
If,when you set up a system using a server 2003 product and do not use the diskpart command when setting up the partitions,you can run into alignment issues which will affect the speed of the SQL server.

Easy reading:


I apologise to those offering advice for not responding sooner.  I have taken on board the advice received thus far and I am in the process of devising a recovery strategy.
This is a first time experience for me (RAID 5 failure) and although I know time is not entirely on my side I will make changes when I feel comfortable with my understanding of the processes.

I emailed this thread to the vendors of the file-server who have subsequently intimated some advice/assistance but none received to-date.

For the record I need to correct an earlier post where I said that we had previously used SBS Server 2003 under RAID 0 control, in fact it was RAID 1 - mirrored.  I knew that all along but I must have got excited when typing!
RAID 1 ,when using desktop drives is a lot more stable than RAID 5.

RAID 5 has a lot of advantages,but you need to understand the caveats.

As for copying to another disk,you could try and see if that works for you.

It's free.


As a tempory expedient I have taken the advice proferred by willcomp and I have located and purchased a 500gb disk identical to the other 2 in the array.  I fitted the new disk late this afternoon (UK time) and the server booted OK; so far so good.

On the array controller the new disc showed as having 512bytes per physical sector, as per the Port 0 and Port 1 discs, and therefore I instructed the array controller to include the new disc in the array.  As I type this post the process is about 12% complete in 3 hrs.  So far so good again.

When the array recovery is complete I will reboot the server at lunchtime Friday to ensure that I have a working system.

Assuming that I will have bought back the time to think things through at a more liesurely pace the next step will be to upgrade to ES grade discs per the advice and software solution given by pgm554.  I guess pgm554 would prefer a clean install of the OS software and SQL software followed by restoration of data and as a last resort it may come to that but regrettably downtime is very limited and less time consuming solutions might have to prevail.

I will keep you posted on progress tomorrow.  But in anycase thank you both for your support and patience thus far.
You may be able to get away with the xml product because you need to have a partition all ready created to copy the image back to.


The Raid 5 array has completed its recovery and has 3 discs back in operation.  

As explained earlier I am now working on a strategy to upgrade the physical discs to ES.

Thank you for your assistance.

Glad you got it recovered.

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial