[REQUEST] Repartitioning a running system disk

[mdownes]

I need some advice on repartitioning a 2008 R2 server’s system disk while it’s running, remotely. The system partition had completely run out of space, but I freed up 8.5Gb, mainly by thinning out the Winsxs folder. I’m not even sure a remote, running repartitioning is a good idea, given that if something goes wrong, the location is 2 hours away. It’s a business which relies on a trade-specific, server-based application. The app vendor provides their IT support, but is no longer doing onsite work. They’ve offered to pay for 2 hours labour for another tech to do the repartitioning. I’ve suggested they get someone local but they can’t find anyone. They’re also unwilling to move to Server 2016 for the next 18 months.

I’ve never done this before, but my first instinct is to get them to connect a USB disk and start with an image of the whole disk, using Easeus ToDo or something similar. Next, I’m thinking copy the contents of D: to the same disk, delete D: partition, enlarge C:, recreate a smaller D:, repopulate it with its data and the job’s done. Am I being realistic or is this highly idealised thinking? Should I use something like Aomei Partition Assistant Server? Or is this too risky (in terms of labour overruns if nothing else) to undertake to do remotely and within 2 hours?

 

[YeOldeStonecat]

I've done this remotely...yes its an uneasy feeling....you're screwing with a servers system volume. When you apply settings and perform the reboot, it's a nail biting moment of time until the server comes back up again.

Is this Server Standard? Or SBS? What size C volume?

I've used two products in the past for doing this, Acronis Disk Director...it's around 600 bucks, and some product by Paragon Software..I can't find its exact name, it was roughly the same price.

First question...what kind of backup do they currently have? I would not touch this job unless there was a full disk image backup system with known good backups. If something goes "oops" on the reboot, you want to be able to roll back to where you started with ease. Meaning the backup system needs to work restoring to bare metal on that hardware.

Second question..because how you answer this may reveal a different approach to your issue. I've used these methods quite a few times also. What kind of hard drive setup is in this server? RAID? How many drives? How many RAID volumes? What type of RAID?
Say this server just has a pair of drives setup in RAID 1. And that volume is partitioned into its current C and D. Say this servers drive bays and RAID controller support adding 2x more drives in there. Slap in 2x more drives, create a second RAID1 volume. Do a backup of the current D partition (second partition of volume 1)...restore it to this new second RAID volume. Confirm its valid. Delete that old D partition on the fist RAID1 volume, assign the second RAID1 volume drive letter D. Now stretch the C partition to utilize the remaining free space of the original first RAID1 volume. So now the server is running all its C drive on the first RAID1 volume, and the D drive is on the second RAID1 volume. Should perform better too. Also toss some of the pagefile.sys over to the D volume...spanning pagefile across multiple spindles gives more performance.

Celebrate with a few pints of Beamish!