V4 requires private key to move EFS encrypted files

[usch]

I just upgraded from JkDefrag to MyDefrag. It's undoubtedly an improvement. ;-)

However, one of the first things I noticed was that it does no longer defrag or move EFS encrypted files, unless the user who is running MyDefrag has a private key to access that file.

To reproduce:

• Log in as Administrator
• Create two files, TEST1 and TEST2, using MyFragmenter.exe
• Make sure all users have read and write access to these files
• Encrypt TEST1 from the Administrator account
• Log in as another user
• Encrypt TEST2 from that account
• Log in as Administrator again
• Run MyDefrag

Result: File TEST1 is processed properly, but file TEST2 is marked as unreadable and remains fragmented.

Now run JkDefrag (V 3.3.4, if that matters) on the same drive: TEST2 is defragmented, too, as expected.

Since this used to work before, I'm sure it is easy to fix. Thanks

[peters40]

Thanks for that usch, it explains why some of my encripted files are not being processed. I have 3 accounts home user, work user , and admin account, now i understand.

[jeroen]

Result: File TEST1 is processed properly, but file TEST2 is marked as unreadable and remains fragmented.

Thank you very much for the excellent instructions. I have managed to reproduce the problem and have found the reason. JkDefrag opened encrypted files just like any other file, MyDefrag uses the method recommended by Microsoft, which is different than for regular files. I never tested with encrypted files by other userids, using only a single userid on my desktop, so I never encountered the problem myself. Anyway, I am scrapping the method that Microsoft is recommending and going back to the good old JkDefrag way of opening files. In other words, the problem will be fixed in the next release.

[boco]

How 'bout using the MS method, but fallback to the old behavior if it fails.

[usch]

In other words, the problem will be fixed in the next release.

Thanks for the fast response!
I forgot to mention that I am running XP 32bit; other Windows versions might of course behave differently.

[cg]

Thank you very much for the excellent instructions. I have managed to reproduce the problem and have found the reason. JkDefrag opened encrypted files just like any other file, MyDefrag uses the method recommended by Microsoft, which is different than for regular files. I never tested with encrypted files by other userids, using only a single userid on my desktop, so I never encountered the problem myself. Anyway, I am scrapping the method that Microsoft is recommending and going back to the good old JkDefrag way of opening files. In other words, the problem will be fixed in the next release.

I was one of the people with difficulty using Encrypted Files  (See the 1.5 GB Memory Post - http://www.mydefrag.com/forum/index.php?topic=1622.0)   When I turned the EFS encryption off, my memory usage dropped significantly and SlowOptimize worked fine.  When it was on, I couldn't complete a SlowOptimize - it eventually used up all my free RAM and just seemed to hang.  I wonder if the so called "Memory Leak" concerns aren't MyDefrag, but using Microsoft's recommended way of opening the files?  Wouldn't that be funny if this fixed those problems as well?

I would suggest using the recommended way only if the JKDefrag way doesn't work - especially if the side effect does actually address the memory concern.

[jeroen]

When I turned the EFS encryption off, my memory usage dropped significantly

Well, I spend a couple of hours on it this afternoon and it all seems to work great now. You're talking about the beta versions in that other thread, lot's of changes since then.

[cg]

Well, I spend a couple of hours on it this afternoon and it all seems to work great now. You're talking about the beta versions in that other thread, lot's of changes since then.

Yes - but I've been using every version up to the 4.1.1.  I haven't done an EFS on/off check on each and every one because of the time that takes, but I know the problem was still there in v4.0 final.  And the change you're making may not make a memory difference, but it is interesting that you found this issue which at least seems to corroborate what I was seeing when I did my own testing.  In my case when I removed the EFS, the memory utilization dropped.  That would seem to indicate that maybe when you use the Microsoft method and it isn't able to move a file, that perhaps the open attempt is using memory and not releasing it.   Again - I could be wrong. 

In any case it sounds like you've got a workable solution for EFS which is great!