Avatar of wheeling
 asked on

Swap hard drives?

I have a dell poweredge 2900 with 2 hot swap hard drives raid 1 for my c drive and 3 raid 5 drives for my d drive.  I am running SBS2008 and SQL server 2008 with most of my data on my d drive.  I am running out of space on my c drive and was wondering if I removed both my c drives and installed larger drives and reinstalled the operating system:  If I didn't have enough time to finish the new drives could I reinsert the original drives and would the computer recognize them with no change?  If so I could later continue my new setup and would the new c drives recognize the d drives?  Also, I have perc6i, SAS RAID controller

Any comments?
Microsoft Legacy OSServer HardwareMicrosoft SQL Server

Avatar of undefined
Last Comment

8/22/2022 - Mon

The Raid information is stored on the 'hdd's' it self. I usually remove the HDD's out label them based on the bay they were in before sending it out for repairs etc.

I tried this once and had no issues. Was in a similar situation as you. Needed to upgrade the root. But since I have issues with the installation, I just popped in the old HDD in THE SAME ORDER and it worked fine.

The only difference was I had Raid 5 (3 HDD's) for the root as well.

Hope this helps.

Easiest swap will be done by imaging your current primary to the new larger RAID 1 array. The two you remove can always be put back (one or both) and booted. You'll save a lot of hours by imaging though.


I am not familiar with how to image my current primary drive.  Can you steer me to a tutorial that I could learn how to?

Your help has saved me hundreds of hours of internet surfing.

View this solution by signing up for a free trial.
Members can start a 7-Day free trial and enjoy unlimited access to the platform.
See Pricing Options
Start Free Trial
Ask your own question & get feedback from real experts
Find out why thousands trust the EE community with their toughest problems.

If you have C: on RAID-1 partition then you might look at the option of using the RAID to your advantage. I have used the following procedure to upgrade a hardware RAID-1 partition:

1. Remove ONE of your RAID-1 disks. This will cause the array to be in degraded mode (ie. not redundant)
2. Put your NEW bigger disk in the slot that you just pulled the disk out of.
3. Use the RAID utilities to re-create the mirror to your new disk.

At this point you should have a functioning RAID-1 array again. It will be over two disks and one (the new one) will be larger than the other and contain free space.

4. Remove the second OLD disk. At this point you will be running a degraded array again (ie. only one disk)
5. Using your RAID utility, expand the size of the ARRAY on the new large disk to take up all of the space.
6. Using a partition utility, expand the size of the PARTITION to take up the full size of the array you have just extended.
7. Install your second new large disk.
8. Using RAID utility re-create the mirror so that it is now over both of the new larger disks.
9. Sit back, have a drink.

The advantages of doing this are that you should be able to do it all "live" with maybe just a reboot to make windows recognise new partition sizes (depends on windows & RAID drivers probably). You also have a fallback method where you haven't trashed the old disks and so that if anything goes wrong you can just put them back in (ie. the very first disk you pulled out that you have changed nothing on). I have used this method previously with good success rates. Obviously with anything make sure you have a backup of anything important...

The problem with td's method is that after you have rebuilt both larger drives, you will be left with an array that yields the same disk space as before.  You'll need the drink that he recommends to cope with the idea that the PERC controllers do not support auto-expansion.  At this point, you would really only have one option:  Create a new array (a new drive letter/partition) across the disks using the additional space.  You might see or be told that you can delete/recreate the array in the BIOS utility for the PERC without initializing it, but this is dangerous and will create an an unstable array.  So if you value your data, then backup your array, delete it, recreate array with larger disks, then restore data to the now larger array - or simply create a new array across the disks to utilize the space.

As you suggested in your original post - yes, it is possible to take out the original disks, reinstall on the new larger disks (or better yet - image old RAID 1 to new RAID 1 as has been suggested by SemperWiFi), THEN if needed you could put the original drives back in and they should work fine.  If you do this, upon putting the original drives back in, they will show as FOREIGN in the controller ... you would simply Import the Foreign Config in the CTRL-R BIOS utility (F2 on the controller, Foreign, Import).

Fair call, I wasn't aware of this limitation for that particular controller. I've used the above method on both IBM & HP servers in the past, but yes it will be dependant on the controller's ability to extend the array.

Best next option would be to go with the other method suggested which is to image your existing drive/volume, remove old disks, install new disks, create new array and then image back to your new array.
Try out a week of full access for free.
Find out why thousands trust the EE community with their toughest problems.

Just image over and be done, takes less time than a lengthy RAID rebuild as well.

Great input, thanks all for your help!  I will come back if I have problems with the procedure