Jump to content
  • 0

Upgrading from WHS2011 to Windows 10 and keeping Drivepool alive.


GaPony

Question

I'm sure this has been asked about a million times, but probably not lately. Allow me to layout my scenario.

I'm currently running WHS2011 on a Supermicro X9SCM-F mobo with an Intel Core i-3 2120T and 16GB Non-ECC RAM. It has run flawlessly since installation back in 2011. I figure when Google web browser tells me its time to upgrade my OS, it's probably getting a bit too long in the tooth.

I'm using Stablebit Scanner and Stablebit Drivepool with a pool of 27 hard drives (connected to 3 Supermicro HBAs) for a total of 120TB of data, along with a backup drive for Acronis. The sole purpose of this server is to feed movies and TV shows to my Plex server on another machine. I have already purchased a 16-port and 8-port HBA as a routine upgrade.

I'd like to upgrade the OS to Windows 10. I'd also like to do it as an upgrade instead of a bare metal replacement for reasons I can really articulate. It's sort of one of those "If it ain't broke, don't fix it (too much) things. Is this possible?

I'm pretty sure I can do a bare metal installation of Windows 10, then install Stablebit Scanner and Drivepool and reactivate my license, pretty easily since I have very other programs installed on this server, except for a few easily replaced utilities. Now for my main concern,

When I originally set up my server, I had substantially more than 26 hard drives, so I instead of trying to use drive letters, I set them up to a folder on my C: Drive (system). They are all linked to separate folders in C:\DM folder. I'm not sure how I would deal with that. Could I just offload the C:\DM folder to a USB drive, to the bare metal installation of Windows 10, then copy the C:\DM folder back? I haven't found any helpful information about doing it. I would feeling like shooting myself (maybe in the foot) if I ended up trashing things.

Sometime over the next year I plan to upgrade all the hardware in order to run Windows 11. It seems to be getting harder to find a new motherboard that supports at least 2 PCIe slots supporting at least 8 lanes each, without breaking the bank.

As I said, I'm sure this general question has been asked a million times, but I feel there are a few unique features to my setup that I don't want to screw up and may be worth someone's attention.

Any help would surely be appreciated.

GaPony

Columbus, GA, USA

 

 

 

 

Link to comment
Share on other sites

7 answers to this question

Recommended Posts

  • 0

I'm not aware of any way to upgrade (rather than replace) WHS2011 to W10.

If you have a folder - e.g. C:\DM - in which you have junctions to various drives in your pool (e.g. due to using Disk Management to mount the paths to the volumes) and wish to copy those junctions to another drive - e.g. E:\DM - then one option is to use the following command to copy the junctions and anything else in the folder (bug: you may need to run it multiple times until all links appear) in an administrator prompt: robocopy C:\DM E:\DM /mir /sj /sl

Note however that the above method will not update Disk Management with the copied junctions as mounted paths; if you want Disk Management to show the junctions as mounted paths then I suggest instead using the mountvol command to list all current volumes and their mounts, then use that output to create a batch file using mkdir and mountvol with the appropriate parameters to recreate those mount paths on the new system. You will need to run this as an Administrator.

Link to comment
Share on other sites

  • 0

Thanks for your reply Shane,

It didn't like that robocopy command. Let me ask my question in a different way.

If I just run mountvol to make a list of the drives and their associated links in the C:\DM folder as a reference, will the system retain the drive identification for the disks (disk 0, disk 1, disk 2, etc)? If those remain constant then manually mounting them back to the C:\DM folder should only take a couple seconds each Though a batch file may be much faster, I would have better control doing it one by one. 

My singular fear is that the system would arbitrarily change things (ie. Disk 5 would become Disk 7 and so on). So long as that stuff remains constant, re-associating the particular drive to the desired path should be easy. I'm not sure how much it would even matter except that I name the folders for the drive's location within the server chassis (ie. Bay_01-01, Bay_ 01-02, etc.). It makes it a lot easier when changing out drives, or investigating S.M.A.R.T errors from Stablebit Scanner. 

Amazingly, I've never had a drive failure since my last major hardware upgrade in 2014, though I've replaced smaller drives with larger ones over time. I still have several of the original HGST 4TB drives running smoothly and error free.

I've made 4 clones of the WHS2011 system drive, so I have some room for screw ups. :)

Thanks again!

Link to comment
Share on other sites

  • 0

Yeah, I forgot robocopy has gotten some updates since 2011, that might be why?

Disk Management's disk numbering does not have guaranteed persistence across OS or hardware changes (and references the physical storage devices, which can each contain zero or more volumes).

Instead mountvol uses volume identifiers - in the format of \\?\Volume{dashed-alphanumeric-string}\ - that do have persistence even across different machines (as the volume identifiers are created with and stored in the volumes themselves).

So if your current tracking method involves Disk # then I would suggest using Volume ID (and/or your own custom unique Volume Label) instead for the changeover.

Link to comment
Share on other sites

  • 0

Thanks for that. I figure the worst that can happen is that I'll lose tracking of the individual drives for a bit, or at the very worst, I'll have to allow Drivepool to rebuild everything and reorganize to my liking later. After more thought, I'm at least confident the data itself is safe enough, though I'm a little less certain about the duplication. I'm guessing Drivepool will figure that out after I reconnect the drives and re-register. I'm not so concerned about the disk #s themselves, it just let's me know where a disk resides in the server. I could always move them around afterwards.  Maybe I should have done this 7 or 8 years ago before my pool grew so large and I only had a few physical drives to worry about.  WHS2011 with Drivepool has just been so bulletproof, I never gave it much thought.

Link to comment
Share on other sites

  • 0

Another thought, maybe a little crazy since its late here and certainly an inelegant method, I could offload about 12TB of data to a "backup-backup" rig. Then remove 4 small drives to get below the 23 drive threshold for using drive letters. Then I can revert to drive letters to ensure nothing goes wonky with the junctions points. Once things are back up, I can simply change everything back to junction points, then add the new physical drives I used to offload the files to the pool. Something I was going to do later anyway. Like I said, inelegant and time consuming, but time isn't a constraint for me. I'm in no rush, plus I'll be freeing up 2 or 3 bays for even more stuff! Win-Win, if it works...

If this sounds completely crazy just say so. I'm too old to have my feeling hurt by coming up with insane ideas.

Link to comment
Share on other sites

  • 0

DrivePool itself doesn't care about the drive letters or mount paths of the volumes that make up a pool; if you do a bare metal reinstall of Windows and DrivePool then DrivePool should simply detect the poolparts from the previous install and recreate the pool drive.

If you're using externally-powered bays for your pool, I would simply use mountvol to note which volumes are linked to which bay positions, shutdown, power the bays off, do the bare metal install of Windows 10, power them on, remove the drive letters that Windows creates for them, install DrivePool, it should recreate the pool, then you can give it the drive letter you want and recreate the mount paths for the individual volumes.

Link to comment
Share on other sites

  • 0

All my drives are installed inside the server chassis, so I'll just remove the HBAs which will have the same effect. Reading back through many, many old posts on the subject, I believe I've been looking at this all wrong. The actual information Drivepool uses to build and maintain the pool is stored on the pooled drives, not on the system disk, so its just a matter of reinstalling Drivepool and it goes looking for the poolpart folders on the physical drives in order to rebuild the pool. I'm not sure why I've been thinking I needed to worry about what was on the C: Drive, other than that's where I placed the junctions for the drives on my system, but that seems to be more of a housekeeping feature to get around the 26 drive letter limit more than anything to do with Drivepool itself. If this is correct, I believe I'm ready to go.

It would seem that Stablebit Drivepool is quite a feat of engineering.

 

Thanks again Shane.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Answer this question...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...