Jump to content
  • 0

Portability of Pool from One OS to Another and Back


Mick Mickle
 Share

Question

Christopher / Alex,

 

Please comment on the portability aspects of moving a Pool from WHS 2011 to Windows Server 2012 Essentials and back.  I'm posting this question here rather than in a support contact because the general information and guidance can also apply to what other users might want to do.  Basically, for testing or trial purposes, what steps and precautions should be followed to move a full pool of pass-through disks from a VM with one OS to another VM with a different OS and back again without risking data loss on the pool?  I think this would also apply to moving the pool intact between different computers or motherboards or upgrading then reverting OSs.

 

Scenario: As you know from my previous support tickets (640360 & 8209559), I've been having trouble with freezes of my WHS2011 VM (Hyper-V) when the DrivePool service is manually enabled (even running minimal msconfig settings and without the Scanner service). This occurred with the latest v 1.3 and also after I upgraded to the latest release of v 2.  Usually the hang will occur at some point during the measuring stage; for some boots, though, the DrivePool service enters a steady state and presumably executes its management tasks okay.  I know that I can operate this way indefinitely without data risk if I keep in mind what Alex told me that I've pasted into the end of this post for the benefit of all.  

 

But after exhausting much of my hardware and software troubleshooting resources, I'd like to move the full single pool of 7 HDDs (16.7TB) to a VM with Windows Server 2012 Essentials to see if the problem disappears.  If it still exists, the root of the problem will probably by the host hardware or software.  However, I'm not prepared to commit to that server upgrade change from WHS2011 yet, so I want to be able to reattach the hard drives to the WHS2011 VM.  Naturally, I don't want WS2012E to massage the Client Computer Backups folder on the pool and declare the backups corrupt.  (The current native WS2012E Client Computer Backups folder is with its other default server folders on the system C: drive now, so I'm thinking that WS2012E won't automatically try to connect to the similar folder on the DrivePool pool.  Still, . . . . . .)  Here's what I think I need to do.  Please correct me where wrong and let me know what pitfalls are in store:

 

Steps:

 

1. Install DrivePool 2 into WS2012E VM and shut down.

 

2. In Windows Explorer, rename (slightly) Client Computer Backups folder in the DrivePool drive's \ServerFolders in the WHS2011 VM (will cause a pre-defined server folder missing error and stop the Windows Server Client Computer Backup Service) and shut down.  (?? - Not sure this is necessary as a safeguard to keep WS2012E from playing with it after the pool is moved.)

 

3. Remove the physical hard disks from the WHS2011 VM using Hyper-V Manager.

 

4. Add the physical hard disks to the WS2012E VM.

 

5. Start the WS2012E VM.  DrivePool should find the full pool intact.

 

6. Run the WSS Troubleshooter Reset NTSF Permissions on the Pool and Restore DrivePool Shares.  (?? - Not sure this is necessary unless the intent is to keep the pool on the new server, particularly if the plan is to return to the old server.)

 

To return the pool to the previous server:

 

1. Shut down both VMs.

 

2. Remove the physical hard disks from the WS2012E VM.

 

3. Add the physical hard disks to the WHS2011 VM.

 

4. Using Windows Explorer, reverse renaming of Client Computer Backups folder in Step 2 of first move above.

 

5. Run WSS Troubleshooter Reset NTSF Permissions on the Pool if it was done for Step 6 above. 

 

6. You should be back to where you started.

 

__________________________________________________________

 

Using DrivePool without the service: In reference to my statement above that the DrivePool service doesn't have to be run in order to use DrivePool (provided you have sufficient free space on all pool disks), I thought I'd include this clarification that Alex provided me in a support contact concerning data risk  --

 

Q. "What are the risks for regular use (of DrivePool with the DrivePool.Service.exe service disabled) until the service is re-enabled?":

 

A.  •Risks to data integrity:


◦If you have real-time duplication enabled: NONE.
◦Background (or nightly) duplication will not function.

 

      •No pool management.


◦You will not be able to add / remove disks from the pool or perform any kind of pool management.

 

      •Risks to running out of disk space:


◦There is a complicated mathematical model that we use to re-balance your pool to provide maximum available disk space for duplicated files. This will not function with the service disabled.

 

A simple example of this (but it can get much more complicated):

 

◾You add a 100 GB disk (#1) to the pool.
◾You fill that disk up to 90 GB.
◾Then you decide to add another 100 GB disk (#2) to the pool.
◾Your pool is now 200 GB in size with 90 GB used.

BUT, you can only copy 10 GB of new duplicated data onto it because disk #1 only has 10 GB free.


◦So if the service is not running, you risk running out of disk space.

 

      In short, the service is not involved in reading or writing data to or from the pool. But maintenance and administrative tasks will not function.

Link to comment
Share on other sites

20 answers to this question

Recommended Posts

  • 0

I went ahead and pulled the trigger.  So, I guess I was using the above to think out loud.  Looks like no need for step 2, renaming Client Compute Backups folder on the pool.  WS2012E doesn't give a whit about that folder since the pre-defined folder for that OS isn't on the pool to start; it would just be another folder with same name on a different drive.  Also, no problem with access permissions due to an execution of WSS Troubleshooter a couple months ago, which left "Authorized Users" and "Users" in the security settings for the pool's ServerFolders.  (I see you've added the recommendation to remove them in the WSS Troubleshooter instructions to get back full security.)  So once WSS Troubleshooter Reset NTSF Permissions on Pool has been used if you make on other permission changes, access should be open on folders in ServerFolders regardless of where you stick the pool, before or after you move it.

 

So far, both WS2012E and DrivePool (with Checking at about 80%) seem happy.  We'll see how the measuring and balancing go. 

Link to comment
Share on other sites

  • 0

Sorry for not getting back to you sooner! I really did mean to. But it was late at night when I got to your post, and I was too tired to respond.

 

 

As for the pool.... I use a single set of drives for testing DrivePool. A series of 4 VHDx files. And I move them from system to system. That's WHS2011, Windows Vista/7/8 (both x64 and x86), and Server 2008R2, and Server 2012 (R2) (both essentials and standard).

 

So, the pool is VERY portable. Very.

 

 

The only real consideration is NTFS Security permissions.  The "built accounts" will be the same between all Windows systems, so that will be fine ("SYSTEM", "Users", "Administrators", "Authenticated Users", etc. Any "Group" or user account you don't have to create).  However, the individual user accounts are uniquely generated on each system.  You can "return" without any issue though.  

 

Also, Windows Server Solution (WHS2011, and the "Essentials" family) assign "SYSTEM" and "Administrators" to have full control. You don't want to change this part (for simplicity, and to ensure DrivePool works)

 

 

If I didn't cover anything that you needed covered, or need further explanations, don't hesitate to ask.

Link to comment
Share on other sites

  • 0

Hey, no big delay -- you're always on top of things, and I appreciate you taking the time.

 

This is exactly the guidance I was soliciting.

 

It's all pretty smooth, as you said.  But also as you said, the NTFS Security permissions is an issue to deal with. On first "import" of the pool from WHS2011 to Server 2012 Essentials, the Security properties has all of the "Account Unknown" user names from WHS2011.  So it's apparent that they would be reinstated properly upon the pool's "return" to WHS2011.  However, if you allow Server 2012 to touch a folder to share it, either by adding an existing folder as a server folder through the Storage tab of Dashboard or turning the Media Server on in Dashboard (affecting existing Pictures, Music, and Videos folders), I noticed that all the Account Unknown user names are purged.  So I assume that when the pool is returned to WHS2011, those folders will need to be re-shared with intended users.

 

I'll say that this exercise of moving the pool is being fruitful, in that I'm not having any of the problems yet that have been long-term issues with WHS2011 VM freezes using DrivePool and Scanner since last Spring.  So, while it's too early to know for sure, it's likely a software incompatibility or corruption that got introduced over time into the WHS2011 VM rather than host or hardware problem source, or it could be specific to a WHS2011 idiosyncrasy.

 

Thanks, again.

Link to comment
Share on other sites

  • 0

I wasn't actually aware that the "WSS" stuff would wipe out the "unknown" SIDs. But that it does, doesn't really surprise me either.

 

And yes, if they were removed, then they would need to be "reapplied" if you move the disks back.

 

 

 

As for the system issue. Unfortunately, it could be something as simple as "bit rot" that is the issue. And speaking of which, have you ever considered running "sfc /scannow" (or sfc /scannow /verifyonly) to see if there is any issue with the files for hte system? It *may* help.

Link to comment
Share on other sites

  • 0

I did run sfc /scannow and did have some corrupted system files which were fixed. (I actually have a command file to run the System File Checker with a pause at the end in my "+Utilities" Start Menu /All Programs folder, which I always create with my most important and useful utilities for any OS.  But I hadn't run it in many moons.)  However, I will need to get the pool back to WHS2011 to test the difference.

 

I would say that wasn't the cause of the problem with the DrivePool service freezing the VM because the problem is also happening on WS2012E. But, SFC also found and fixed a corrupted file(s) there. Nevertheless, that guest still freezes when DrivePool service is measuring.

Link to comment
Share on other sites

  • 0

That's.... upsetting. It may be just coincidental. But I wouldn't be surprised if there was some underlying problem that DrivePool brought out.

 

Check that system's disks for other errors (such as SMART errors)? And it may be a good idea to run a memory check too. Just in case. 

Link to comment
Share on other sites

  • 0

I think you hit on the situation correctly, DrivePool is bringing out an underlying problem. I just can't figure out what it is. I ran memtest86+ on the 12GB of RAM for about 8 hrs yesterday - no problems. All 9 of my HDDs pass all the tests ok except one which has 7 reallocated sectors in SMART, otherwise stable. Because I had one drive drop out apparently due to a faulty intermittent factory power cable connection, I replaced the PSU with a new one today. Still hangs when I run the DrivePool UI!

 

Please do me favor and bump 8209559. I also uploaded a dump on Nov 24 that I forgot to key to that support contact, and I'm not sure the upload completed.

Link to comment
Share on other sites

  • 0

.... That's a nice long ticket. 

 

And unfortunately, it does look like it's an issue with the HyperV driver. So I'm not sure how much we can do to help here.

 

Though, have you tried removing the Samsung disk from the system completely and see if that helps?

 

Also, you say you have a faulty power cable connection? was this for the HDD? If so.... it could be that it may have damaged the circuit board on the HDD. And that *could* definitely cause this sort of weird/erratic behavior.

Link to comment
Share on other sites

  • 0

I don't think the PSU cable would have damaged the HDD circuit board. Any cable defect would have only been the mid-cable connector may not have been crimped well enough to keep an intermittent open circuit from occurring.  That would have, at worst, been similar to hot plugging a SATA drive.

 

No, I haven't removed the Samsung drive from the pool.  If I do, DrivePool will just complain that the Pool isn't intact, so it won't execute measuring, which is when the freeze occurs.  (And I hate to migrate everything off an apparently good drive for testing.)

 

I do wonder about DrivePool handling of > 255 character paths and whether that's a factor.

Link to comment
Share on other sites

  • 0

That may be true. But you should have "should" instead of would there.... Weirder things have happened....

 

 

As for the samsung drive, you could "remove" the disk after disconnecting it, and you would be fine. Well, it would remeasure and reduplicate as needed. Reconnecting it may "correct" the pool config though. 

 

As for the 255 character paths, I do believe it does support that just fine, but I'm not entirely sure.

Link to comment
Share on other sites

  • 0

Touche on woulda, coulda, shoulda.

 

How would "removing" one drive after unplugging it to get DrivePool running normally affect client backup database? My concern is that the backup service would decide it's corrupted. Of course I could stop the backup service, but then that wouldn't be a full test of DrivePool. (Maybe client computer backups folder isn't on that disk, though. I'll check.)

 

On 255 character path, could you confirm with Alex?

Link to comment
Share on other sites

  • 0

Christopher,

 

As you know from my new support contact, removing the drive didn't help, and the lockup still happened after duplication completed. So, I decided that one thing I haven't tried yet is too move the pool to the host and see if the problem still exists on the host, as opposed to the VMs. However, I've run into a problem with DrivePool not recognizing the existing pool. (I didn't want to complicate the support contact issue, and this subject is a perfect continuation of this forum thread, anyway.)

 

I installed DrivePool 2.0 current release in 30-day trial mode on the WS2008R2 host and put all the pool drives online (previously in offline pass-through mode to Hyper-V VM). But DrivePool doesn't show the pool or create the new pooled drive. I've tried rebooting and resetting DrivePool according to the wiki. And I can't "add" any of the pooled drives to, because the error detail says can't add the drive twice. (So really, it does know there's a pool!)

 

What now?

Link to comment
Share on other sites

  • 0

Works now -- got the pool okay.  It must have been a corrupted installation of DrivePool.  I recall that the installation failed, I had rebooted and successfully installed over the failed installation.  Finally, I did a complete uninstallation, then deleted all remnants on the C drive and in the registry, rebooted, and reinstalled. That did the trick.

Link to comment
Share on other sites

  • 0

Yeah, the first corrupted install was my fault. I keep my software installation files on a shared folder on the Pool. I installed from there and while that was going on I got ahead of myself and removed the pool drives from the vm to make them available for the host. So the installation failed at first. Duh!

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...
 Share

×
×
  • Create New...