1

This is a sanity check and a means of getting some recommendations.

One man shop that inherited a couple modestly equipped DL380 G10s hosting Hyper-V. The 380s are configured identically with an 8-bay cage and 8 port raid controller but their OS (2019) is taking up drives via raid 1. I'm getting ready to order some expansion cards and additional cages but my primary goal is to move the host OS from said drives to an m2 and ultimately allow the data partition to use that spinning rust. The data partitions on both hosts are raid 6 connected to the same controller as the OS logical partition i.e. separate drives, separate logical parts between data and OS.

First thought was to:

  • Decrypt the drives (bitlocker)
  • Power down Hyper-V guests and host
  • Install m2
  • Physically pull drives associated with data partition to prevent accidental format
  • Boot to Windows PE
  • Capture the OS via DISM
  • Apply WIM to the m2 and write bcd files
  • Power down host
  • Remove original OS drives
  • Reinstall drives associated with data partition
  • Boot
  • Confirm functionality and eventually format the original OS drives to give the data part a few terabytes.

I'm leaning this route namely because these host OSs don't have agents running on them yet, they're somewhat recently built, yet live in an air-gapped environment so patches and hardening is done manually for the time being (though soon to be resolved). Bringing a fresh OS back up to baseline (times 2) would be a sink I would like to avoid if possible. If you have any recommendations, warnings etc I would very much appreciate it.

Improve this question
0

2 Answers 2

Reset to default
0

Stupid question - why not reinstall?

You say " a couple of" - that should be enough to move all stuff off one of them, otherwise they are reckless anyway (servers fail). Free one, then reinstall as wanted. Prepare a script for that (not necessarily computer - a checklist may be faster) and good install media and you can do one an hour. Talk to them - that may well also be the moment to upgrade to 2022 (it is not like 2022 does not have a lot of better features actually).

Improve this answer
1
  • Stupid question - why not reinstall? Not a stupid question at all.. I posed the same and it was determined either host of the two does not have the resources to handle the entirely of their VMs. I've since ordered them a third host which will likely serve as fail-over for both of the existing boxes. Its small company that deals with Gov contracts, think SCAP\STIGs, so not as swift as a standard bare metal build but not a killer task at the end of the day. For that reason they're also stuck with 2019 until 2022 gets approved and procured. Thanks for the input
    – AKBingo
    Oct 9 at 21:51
0

I would also create a backup before hand. And you also need to update the boot configuration at some point to boot from the new drive.. You could use third party software like clone zila to clone the disk

But since it's raid, the safest thing to do is add the new disk to the RAID, install it using dual boot. Then boot to and configure the raid disk and then install windows in a dual boot scenario Then try to import the VmS. Once you are booted in the new os install hyperv and import the vms

Improve this answer
1
  • Hi Ace, thanks for the input. The first task I undertook was some bare metal backups so we should be all square. It was determined either host of the two do not have the resources to handle the entirely of their VMs. I've since ordered them a third host which will likely serve as fail-over for both of the existing boxes. Near term I may do a variation of what you're suggesting but the hardware raid controller does not like mixing interface types in the same logical partition. The new m.2 drive should be able to take the captured OS image straight from DISM with CloneZilla as an alternative.
    – AKBingo
    Oct 9 at 22:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .