One could learn this the hard way. Given the fact the System restore and File protection (Volume shadow copy for restoring previous versions of files) is announced as safety switch in case something happens on your PC, and it has been the most "popular resolution"
for all ransomware infected and affected users, it is really not fair that this issue has not been either pointed out, announced, or resolved. Instead, giving the users insurance that they are safe for such cases, only to find out the hard way that they've
actually not been protected and that they lost it all, is really wrong and actions should be taken for this!
We've just been running extended tests on our company, with me being the alfa IT specialist, and found out that almost none of the shadow copies on 6 different types and sizes of SSD and 14 HDD, on Windows 7, 8 and 8.1, together tested 63 machines/systems,
(yes, that's what we've done to make as tight conclusions and resolutions as possible) were restored without issues. Most of the systems were clean installed, while few of them older installations therefore having some older restore points as well. Timeframe
3 weeks, giving us enough VSS to try to recover some files as per @drew010 suggestion, we've had some luck, but found it difficult to make it an automatic routine, as it's almost impossible to determine what part of the file has changed since or not, and since
non-changed files are not being shadowed therefore the shadow copy usually contains only one copy of file, it being the one with null data in it.
We've made these tests after dozen of our clients reported that our ransomware recovery resolutions were not successful (victims not having backups elsewere and having VSS on were restored using that one, but files were never checked).
We've also found out that there is probably something about spacing as well. I'll try to describe one example:
HDD: 240GB;
Host OS: Windows 7 Home Premium
Partitions: 2, C:\ 190GB, D:\ 30GB,
Shadow copy enabled: C:\ only
Space used on C:\ : 160GB, 70GB user files
Dedicated maximum space for VSS: 9,5GB
Space in used on the day of restoration, reported by the VSS settings form: 4,2GB
Files affected: all user files on hard drive, ransomware infected test host OS
Previous versions of files available on pretty much all folders
Trying to restore several folders with 200+ MB size: all "successful", all included corrupted/empty .txt, .jpg, .rar files (rar archives being reported with "corrupted header" error, containing only parts of archive, able to extract those...)
Trying to restore folder with 12GB .avi encrypted files: recovery collapsed during process, 2 files being falsely recovered, one being corrupted, and other being partially recovered.
Available previous versions of files: 3 points, (as per 3 restore points stored in host OS at the time).
Main question: Why and how did the OS try to restore files that were bigger than his whole used VSS copy at the time?
Why did the VSS only used half of available space (aka it did not run out of space for this to occur), when it should use all he has available, if even more files were changed?
Why it is reported that shadow copy of the files are available, if they are not really there, as they can't be due to space limit?
Why even though it didn't have enough space for even one single shadow copy, it yet suggested to go as far as 3 versions back with it?
This sound like a seriously unstable situation and feature bug. And being completely M$ oriented company as well as their representative, it's giving me serious concerns as per what we're distributing, as this is the second serious thing we've discovered that
affected lots of our clients.
First being the software RAID (mirror) function inside the disk manager itself, where the mirrored disk gets synced - but in a way that it's completely whole mirrored/overwritten again from the source, after every system crash (aka not proper shutdown), giving
the actual point of mirroring exactly the opposite of what it's been made for, and also giving headaches with worn out hard drives due to constant writing that shouldn't been made (I'm totally aware that system crashes shouldn't be happening either, but then
again, that's not the answer to the main issue.)
I'm really looking forward into someone explaining the VSS issue and space allocation!
Best regards,
MarcS.