[Sithtiger]

Originally Posted by ghayes

Here's what is going on. When defragmenting with PerfectDisk and Smart Placement (default defrag method), PD is not only defragmenting files, it is also consolidating free space, placing boot files, and grouping files together by modification date. When the built-in defragmenter is defragmenting, it is doing far less work. VSS is detecting the drive changes generated by PD as changes and as a result older restore points/shadow copies get purged. If you have PD defragment using the Defragment Only option, PD essentially works like the built-in defragmenter - simply defragmenting fragmented files - no free space consolidation, no boot file placement, no grouping of files.

VSS does have code in it to detect defrag activity and limit purging restore points/shadow copies (MS KB article). However, that code only works if the cluster size on the drive is a multiple of 16k. The default cluster size on Vista is 4k. Why not 16k if it will limit the impact that defragmenting can have on the life our your restore points/shadow copies? Because NTFS compression doesn't work for cluster sizes greater than 4k.

The first time that you defragment with PerfectDisk using Smart Placement will result in the most changes to the drive (as seen by vss) and when it is most likely that restore points/shadow copies may be purged. It all depends on the amount of work that PD is doing, how much total free space there is on the drive, how much space that restore points/shadow copies take, etc... Subsequent defrag passes should result in less changes to the disk and less likelyhood that restore points/shadow copies will be purged.

- Greg/Raxco Software
Microsoft MVP 2003-2007
Windows File Systems

Disclaimer: I work for Raxco Software, the maker of PerfectDisk - a commercial defrag utility, as a systems engineer in the support department.

Thanks for posting here Greg! So, are you saying the more I use PD, the less likely it will delete the restore points then? Will it eventually leave them alone or should we just not use the default Smart Placement mode? Personally, I would rather use Smart Placement since that's part of the reason we use PD, because it defrags better than the default defragger with Vista, but if we have to disable or bypass that then I guess that's what we'll have to do.

So it sounds like your saying MS needs to release a hotfix in order for 3rd party defraggers to leave the restore points alone in Vista so that they can read the 4k clusters. Diskeeper has this problem too BTW. I just don't understand why Raxco can't issue a fix in the way of an update as well. I don't really care who fixes it, as long as it does get fixed at some point fairly soon. Anyway, thanks again Greg.

[ghayes]

Defragmenters aren't touching the restore points and making them go away. This isn't something that Raxco (or any other defragmenter) can fix as it has to do with how restore points/shadow copies are "stored" on the drive, how VSS detects changes to the drive and handles those changes. Programs like GoBack also have this issue.

[Sithtiger]

I know PD has something to do with it because I've tested it and ran nothing else and it does it whenever I run PD. I still have to try to run in within Windows, but if I do an offline defrag, it always does this! Before I ran PD, I had restore points for 9 days I believe.

[Rollin' Rog]

Should the same issue be expected to cross platforms and affect XP in the same way?

I don't believe we have seen any evidence of that.

And why were all but 1 of Sithtiger's dates removed on a much larger drive with substantial free space?

[Sithtiger]

Rollin' Rog, I forget, did you run PD on your Vista machine or on your XP machine or both? Also, are you running Vista x32 or x64?

If you're running x64 like I am, then I wonder if it's a 64-bit issue, but it shouldn't be because I installed the 64-bit version of PD. Thanks for your help Rog, I'm about done with this though. I've found the problem and I still need to run Vista's defrag to see if it does the same thing. I doubt very much that it does, but I'd like to see just the same.

*EDIT*
Hey, I just thought of something...I don't know if this would work or not and if it does, it's a little annoying, but I'd like to know. Could you burn some restore points to a DVD-RW and then run a boot time defrag of PD and then simply copy the restore points back after it's done? Also, if you could do that, where are the restore points located and are there just files to back up or are there a whole list of files needed to do so. Sorry, I don't feel like looking right now, but if ya know or if someone else does, I'm all ears.

Also, I ran PD in Windows and it didn't delete any restore points. There is a down side to this though.....by doing it within Windows, it took and hour and sixteen minutes to defrag the whole drive as opposed to around 5 minutes the last time I did a boot time defrag. I'm guessing that the boot time defrag didn't defrag the whole HD, but only fragmented parts. The defrag I performed in Windows defragged the whole drive though. That's why I was wondering if one could just burn some selected restore points onto a DVD and then just restore them later. If this only involved burning a few files, it would be worth it since it only takes a few minutes to do a boot time defrag vs a Windows defrag that takes a long time to perform. BTW, PD said I didn't need to defrag my HD and it still took all that time. I wonder if all 465GB were full, or at least most of it, if it would take the same amount of time that it did tonight. If it does, then defragging 465GB in and hour and sixteen minutes is REALLY good.

The whole reason I like the PD is because of the boot time defrag. It's much faster than Diskeeper is, but again, DK suffers from the same problem PD does and deleted restore points as well.

[Rollin' Rog]

I ran it in 32 bit vista within Windows on a 40 gig C drive. I have not recently run it on XP.

I don't think there is any practical way you can backup restore points specifically in Vista.

Lol, here's a follow-on to our thread here

http://www.eggheadcafe.com/software/...eleting-r.aspx

[Sithtiger]

Heh, funny.....maybe it will spread around the net (no pun intended to AotS) and more people will complain to Raxco about it, causing them to fix the problem.