PowerEdge 1950 Perc5/i Raid Increasing HD Capacity

Hi there!

We have a Dell  PowerEdge 1950 server with Perc 5/i RAID controller running Windows 2008. Our box has two Hard disk 146GB each. We want to upgrade the size of the HD to 1TB each.

From the other posts I understand that first I need to break the raid, put new HD in the new slot, use a third party software to make a clone of the HD (EaseUS Advanced Server Backup) and then remove the first HD and replace it with the cloned bigger HD.

My question is that, after I take take out 2nd HD from the slot and put the new one, the system is not recognizing the new HD? How I can make the system recognize it? Shall I create a new volume disk first? I am a new bee, if you could let me know detailed process it will be great.

Who is Participating?
mrklaxonConnect With a Mentor Commented:
No slots left to add disks?

In a server like yours, the RAID controller (PERC in Dells) talks to the disks, not the OS.  You need to go into Server Manager (Dell software) to see what the PERC card is seeing.  The new disk won't show to the OS until it has been added to some type of array and made active.

As dlethe pinted out - BACKUP B4 U TOUCH ANYTHING.

The array is degraded at this point but still running.  I've never followed the process you are describing but would bet it's unsupported by Dell.  Once the new array is created on the new disk the OS will see it and you probably can image it but when you then need to get the PERC to make it the boot array, etc.

You could also re-mirror to the new disk but it won't help the OS partition size, it will only mirror to the same start size and then show unpartitioned free space.

If you have Dell support you may want to start with a call.  Some controllers can expand arrays on the fly but that is normaly a RAID5 thing not RAID1 and even then it only appears as free unpartitioned space.

If you follow the steps you outlined I think you may need to leave the drive where it is and use the Dell Manager to set it as the primary boot array.

Never done what you are describing, sounds more like an options for a dumb controler with mirror ran by the OS.  Can't say it will work.
DavidConnect With a Mentor PresidentCommented:
The advice you were given puts your data at extreme risk .. as you have found out.  The proper way to do this is create a full bootable backup, yank the disks, put in replacement disks, build new array, then restore.   Then you can always fall back on your original data in the even that this happens.

From this point, only thing you can do is replace the disk with the original. The controller will sense the system has a degraded lun, then you tell it to rebuild.  THEN you do it the right way.

(Take a full backup now, if you have not already done so.  A rebuild can be stressful and you have the highest probability of a drive failure during the rebuild.  Lose another disk, and your data is gone).
Be sure you know what you are doing in the Array manager.  Don't delete your current array or destroy it creating a new one!  You can have more than one array.

If you get stuck with a dead system - keep your old #2 disk out until you can contact Dell.  They may be able to salvage the old disk config from it and get you back up.
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

I assure you, as a former RAID firmware architect, this is NOT supported and the single worst method you can use to "upgrade" the storage.  Even after you finish the job ... guess what, the LUN will be the exact same size as it was before :(

There will just be an unused partition that the O/S won't even see for all your efforts.
Not that I'm chearing him on but why would it be the same size?  IF he creates a new LUN with the new disk and leaves the old one as is (minus a disk) then wouldn't it be whatever size he sets it to?  The mirror wou have to be broken unless he puts the disk in a new slot otherwise the PERC will want it for the mirror on the original LUN.

To be sure, he will be messing with the existing Array and LUN config and could end up with a mess.
good question.  The LSI firmware handles this as rebuild, so first thing it does is set the usable capacity on the disk to match the replacement.  So if the disk was 12345 blocks before, then the firmware will mark it as 12345 blocks first, and rebuild the data from the existing disks & parity.  Repeat the process for all disks and you end up with all disks having the same number , 12345 blocks.

The controller has no knowledge of the file system, or partitioning.  It won't expand the LUN in any way.  Now SOME controllers have this feature, but they will never do this on a rebuild, which is what you force the controller to do when you swap out disks.
.. sets capacity to match the failed drive, not the replacement, then it lays down the metadata, then it goes into the rebuild logic.
I agree with the matched capacity but if the mirror is broken first then it won't try to rebuild.  You may know if the PERC allows this while keeping the LUN functioning on the remaining disk, I haven't looked into it that far.

Don't know that it allows designating the bootable LUN either as you would need that to boot after the image process is done.  Moving the disks around would be a problem for the controler's LUN config.
I write firmware/diagnostics for the controller family (this is LSI's product).  The firmware lets you do it, but it isn't automatic. If you replace the drive with the original, and haven't done anything to destroy or modify the existing metadata then the firmware is supposed to treat this as if the drive was temporarily offline.  It is supposed to initiate a rebuild.  (But Dell could have turned that off in their modified factory config.  I just know for a fact that as a developer, I can tell the controller to do this).  

Reason this feature is there is to deal with external enclosures.  I.e, if the LUN was split into two enclosures, and one lost power, or if the user messes up shuts down the external enclosures before halting the host PC with the controller, then such a feature insures proper recovery after a power cycle.

So, AFTER you back up, put in the original drive.  If it does NOT start a rebuild, then turn off the PC, and power cycle with the disk installed. This may get it going.  If not, then you'll have to rebuild the array with the new drives anyway.
write2avirAuthor Commented:
Dear dlethe and mrklaxon,

Many thanks for taking out your time to reply to me. Your replied have helped me clear the concept.
Luckily I haven't done any damage to the system. I put back the 2nd HD (old one) and rebuilt it using the RAID BIOS utility. The system is working properly now.

Before I attempt to try it again I need to reconfirm the whole process with you again.

- Make a full bootable backup.
- Take out the old disks.
- Put in bigger capacity disk.
- Built New Array and then restore from old disks


Thanks again.

Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.