HELP!!! My VMs are BROKEN!!!

I upgraded to Parallels 10, and used it like normal. Everything was fine until I used the functions to save disk space. Now the following issues have cropped up:

My Windows 7 VM is stuck in an infinite system repair loop. (I NEED THIS ONE!!)
My Windows 8.1 VM is stuck on a black screen, it won't even boot to the lock screen.

Only my Windows XP VM still works (outdated thing I keep around for...reasons).

Assistance in fixing these issues would be appreciated. FYI, I have NO BACKUPS!!

Parallels 9 didn't break my VMs on disk space optimisation, so I didn't have reason to worry (or so I thought).

I can't even find the right expletives to describe my feelings right now.

NEVER TOUCHING RECLAIM DISK SPACE AGAIN!!!

 

I was forced to reinstall Windows 7, and it now works albeit with an intolerably oversized disk image. Windows 8.1 is still having issues. Attempts at repair throw up the illogical error "The drive where Windows is installed is locked. Unlock the drive and try again." I have no idea what to do, I upgraded to 8.1 through the Windows Store, so the only disc I have is from Windows 8, and a full reinstall may not be possible as it's an upgrade disc.

Additionally, since I've discovered the cause to be the new implementation of the space reclamation function, I have ceased using the function which is causing the VMs to inflate to intolerable sizes with empty data. I have many other large files I must have the space available for.

If it comes to it I may be forced to buy an external disk to house the VMs for the time being.

 

Yes, looks like Reclaim Disk Space -operation can corrupt virtual disk images. It happened to me twice before I realized why. It's not merely your Windows OS being broken but potentially all the data in the image may be affected. There are also other reports in other threads in this forum for the same problem. I have already reported this as a bug but haven't gotten any reply from Parallels. Unfortunately my observation is that you really need the backups now... I think this is VERY SERIOUS BUG in PD10, which CORRUPTS CUSTOMER DATA PERMANENTLY. And it should be fixed with an emergency fix ASAP before many others are going to lose their data too.

 

BeebarbX - I will open a support ticket for you to investigate why it happened. Check your email in a few hours.

 

I just read the support ticket e-mail, and one thing strikes me as alarming. Why do you need a copy of my VM? I don't feel comfortable doing so, as I could have personal information in any one of them.

Though suppose I could recreate the problem with one of those modern.ie trial VMs instead of my existing ones, as I just got them running again by reinstalling Windows, and don't feel like breaking them on purpose just to create a problem report.

Surely this is an issue that can be recreated by anyone? There's nothing special about my machine, it's just a late 2012 21.5" iMac running Mavericks.

Additionally, though I may recreate the issue in order to make a problem report at some point, I'm not sure how that would help. The problem seems to be that the Reclaim Disk Space function is deleting more than it is supposed to, potentially corrupting the Windows installations and destroying user data. A problem report seems to collect general system information and information on the state of the VM while it is running.

If the disk image is already compromised before the VM is running again, a problem report made at that point may only show a VM that is running normally. This does not seem like an issue that can be isolated with end-user supplied information.

 

I tried recreating the issue with one of the modern.ie VMs, and it didn't happen. So I'm thinking it may only affect pre-existing (from PD8 or 9) VMs, or there must be something with the disk images.
The disk images in my existing VMs are dynamic (split), and the modern.ie VM was just an ordinary dynamic disk image. Perhaps Parallels 10 has issues compressing split images?

I will convert my VMs to plain dynamic disks for the time being. If the problem re-occurs at any time I will try to make a problem report.
Apologies but I'm no expert, and I am not willing to destroy my VMs just to get the diagnostic data you may require.

 

I just tried the reclaim disk space function on my Windows 7 VM after converting it to a regular dynamic disk. Whatever the specific issue is, I am certain it may only affect those using split dynamic images. Though it makes defragmenting the files for the disk images a little inconvenient (2GB chunks being easier to deal with than one large file), I believe I will be converting all my VM images to standard dynamic disks to bypass this issue for now.

So it seems it is not an urgent fix, but I advise against using split images for the time being.

 

I just had to chime in and say that this hit me as well. I lost three different virtual machines to this bug. I was also using dynamic split disk images. I honestly can't believe Parallels shipped this product with this bug in it. As a developer (OS kernel and hypervisor dev) myself I understand that bugs happen. But I am stunned that the virtual disk function was not tested better than this.

 

I think this happened to me as well but I cannot pinpoint the issue. I upgraded to parallels 10, yosemite beta 2 and reclaimed hard disk space all in same day. When I did that my virtual machine crawled to snails pace. The weird thing is it affected other virtual programs as well like VMware. I installed vmware and went out and bought a new windows 8 pro edition and installed on vmware and same issue. Just so slow. I have no idea what to do except try a complete clean install and format my drive with yosemite again. My mac is unaffected but my virtual machine has crawled to a halt no matter what virtualization software I use. I am at a loss.