Rollback Rx-Fragmentation Problem

[Supersnake]

Originally Posted by bgoodman4

Sounds reasonable but the publishers recommend not running 3rd party defragers once the program has been installed. I think it may have something to do with the defrager trying to defrag snapshots but I could be wrong about this. I would suggest that if the publishes recommends installing the program if you are going to do a defrag, and the re-installing it once the defrag is done, that there must be a good reason to do so. I think (and I am just guessing) that once RollBack is installed you are always running inside a snapshot. For some reason I have gotten this impression from reading different articles on the publishers website. If this is in fact the case then it might make more sense that a 3rd party defrager might have a problem with doing its job. Regardless there is a defrager within RollBack that the publisher says will defrag snapshots (perhaps this is why I have the impression that the system is running in a snapshot, it does not make a lot of sense to me to defrag a backup, you would only defrag an active drive ---- I think).

Does this make any sense at all?

Yes, what you wrote does make sense and the consensus of information that you have gathered has convinced me to not run a 3rd party defragger after RollBack is installed. Thanks.

Looks like I am going to avoid RollBack then since I don't want to give up my use of Diskeeper 2009 Pro Premier which is providing me with stellar defragging performance.

Regarding RollBack's method of defragging, i.e. defragging snapshots.
It is files that become fragmented and I want my files to be defragged not my sector snapshots. Sounds to me like my files are going to remain fragmented after RollBack "defrags'. And the only reason RollBack is not defragging "files" is for the reasons that were discussed before.

[bgoodman4]

I don't think thats the case, I think the current state is in effect a snapshot of the entire PC at that time and that when you add a snapshot to the archive only the differences between the older snapshot and the current state are contained in the newly created archive file. Its an interesting concept if I have understood whats happening correctly (which may well not be the case at all). I think this may explain why restoring to an earlier state takes such little time. In effect all that has to be changed is the difference between the current state and the older one, in effect a reverse incremental. If that is the case then when you run a RollBack defrag any fragmented file of whatever nature will be dealt with.

I get what you are saying about Diskeeper as I have been using it for years but what I would lose by not using RollBack far outweighs what I gain from Diskeeper. It really is not a big deal to defrag manually say weekly using RollBacks defrager, I doubt that I will even notice the difference. In fact the main reason I went to Diskeeper in the first place was because I was using Norton GoBack and a normal defrag schedule caused GoBack to lose its restore history. The fact that the PC was continuously defraging on the fly meant this was no longer an issue. Now that I am finally getting a GoBack replacement on the fly defraging is, for my purposes, not very important. Of course we all have our priorities so there is no correct way but only whats right for the individual. For me its the assurance that my data is safe thats key. With RollBacks ability to store up to 60,000 snapshots on 2 or 3 percent of the drive (the idea still boggles my mind) means that I can set the program to update the archives hourly without choking my drive. I am also particularly fascinated by the ability to restore to the last snap and then recover data files that have changed since the restored to snap. Really quite the program if it works as advertised and from what folks on this forum are saying I gather its not just marketing hype.

I know I sound like a shill for the company but I assure you I am not.

[Supersnake]

Originally Posted by bgoodman4

In effect all that has to be changed is the difference between the current state and the older one, in effect a reverse incremental....(cut)

If that is the case then when you run a RollBack defrag any fragmented file of whatever nature will be dealt with. (cut)

I know I sound like a shill for the company but I assure you I am not.

Sound like a shill? Not at all, I can sense your genuine enthusiasm.

I too used GoBack back in the WinME days and I fawned admiringly over it. However, there were times then when I was experiencing BSOD's and if it were not for my insistent friend repeatedly telling me that GoBack was the responsible element for the crashes (because his mother also was experiencing the same complaint while using it) and that I should try uninstalling it to see if my BSOD's resolved like his mothers' system resolved after the uninstall - then I would still be using it. Am getting side tracked here but I did have a love hate relationship with GoBack - I won't even get into the issues with watching my HD running what seemed incessantly and the not so unnoticeable slowdowns while GoBack was logging and archiving which I pretended to not notice. Getting long winded here...

THEN along comes RollBack with a novel way to trump GoBack by taking "sector screen shots" instead of massively transporting files from one place to another on my HD as GoBack does - and my romance is about to resume again - this time with Rollback. But I will have to be convinced that a RollBack defrag really defrags.

What I would like to see is if someone would perform a RollBack defrag on a Hard Drive that is fragmented then uninstall RollBack and have Diskeeper analyze the disk to see just how well Rollback performed a defrag.
I presume that in order to run this challenge one would have to have Diskeeper uninstalled first before RollBack defrags?

Sorry for making a big deal out of this, yes I could just tell myself that defragging accuracy doesn't deserve the priority that a well executed RollBack would provide if I get into trouble.

[pandlouk]

Hello guys,

I have thoroughly tested Rollback-RX/Eaz-Fix the last two years and I can answer most of the questions about fragmention and Rollback-RX.

- Rollback RX uses a hidden file system, which file table is stored versus the end of the last protected partition. If you open your disk with a hex editor you will see that the file table is very similar to the HFS.

- A snapshot is nothing more than a single file of the Rollback RX file system.

- Rollback native defrag does nothing more and nothing less, than to release the free space from the snapshots and fill the gaps between their internal free space. That is why is very important to defrag Rolback RX from it's preboot console every now and then. When you are in the windows enviroment the os also has access to the free space and rollback cannot fill those free gaps in a thorough way.

- The only snapshot that Rollback and windows, never touch is the baseline shield.

If you can understand the above you will see that Rollback RX causes more fragmentation to the windows file system and will not allow normal defragmenters to run because by using them, the only thing that you achieve, is to defragment the current snapshot but at the same time move around all the other snapshots and causing fragmentation to its own file system. If you revert to a previous snapshot you will see a big degradation in the overall speed.

ps. The corruption that occured (and it was no myth) when a chkdsk run was caused, probably, because rollbackRX driver was not loading soon enough and windows were seeing some parts of the original mft and some parts of the mft stored in the current snapshot.

hope it helps,
Panagiotis

[bgoodman4]

You are talking at least a bit over my head (maybe more than a bit at times) but I think I get the idea. A few questions

Am I wrong that Rx is running inside a snapshot? Sure sounds like it from what you wrote.

If I understand you correctly the added fragmentation caused by running RollBacks defrager is not caused if run pre-boot, is this correct?

If there is a problem using 3rd party defrag programs because they try to defrag the snapshots would it not be possible to created a partition on the drive and tell RollBack to store its snap shots there, then just have the 3rd party defrager not do anything to that partition?

or,

Is the preferred route to periodically uninstall Rx, do a defrag, and then reinstall it?

[Peter2150]

Originally Posted by bgoodman4

You are talking at least a bit over my head (maybe more than a bit at times) but I think I get the idea. A few questions

Am I wrong that Rx is running inside a snapshot? Sure sounds like it from what you wrote.

If I understand you correctly the added fragmentation caused by running RollBacks defrager is not caused if run pre-boot, is this correct?

If there is a problem using 3rd party defrag programs because they try to defrag the snapshots would it not be possible to created a partition on the drive and tell RollBack to store its snap shots there, then just have the 3rd party defrager not do anything to that partition?

or,

Is the preferred route to periodically uninstall Rx, do a defrag, and then reinstall it?

Rollback runs in the baseline snapshot where it was installed, and thus will be in all additional snapshots.

You can't store snapshots elsewhere, as there is nothing really to store. As Panagiotis wrote, when you take a snapshot, you have only created a new table, that points to sectors. Many of the files will be in the previous snapshot, new stuff is write as sectors and the table points to them. There are not seen by the windows file system.

You need to take some time and get your mind wrapped around Rollback works to use it most effectively.

Pete

[pandlouk]

Originally Posted by bgoodman4

You are talking at least a bit over my head (maybe more than a bit at times) but I think I get the idea. A few questions

Am I wrong that Rx is running inside a snapshot? Sure sounds like it from what you wrote.

RX resides in three parts of the disk:
1. the baseline shield and subsequnetly in the snapshots
2. In the mbr
3. in it's hidden file table.

Originally Posted by bgoodman4

If I understand you correctly the added fragmentation caused by running RollBacks defrager is not caused if run pre-boot, is this correct?

Wrong. Is caused anyway (I'll explain it further on...)

Originally Posted by bgoodman4

If there is a problem using 3rd party defrag programs because they try to defrag the snapshots would it not be possible to created a partition on the drive and tell RollBack to store its snap shots there, then just have the 3rd party defrager not do anything to that partition?

As Peter already explained this is not possible because of the nature of RX.

Originally Posted by bgoodman4

Is the preferred route to periodically uninstall Rx, do a defrag, and then reinstall it?

Yes, once a month (it depends on how you use the pc). You should either uninstall or revert to the baseline.

For explaining a little better I'll make an example. Let's paragonise RX with an imaging program. In this chase:
- the baseline would be the principal image.
- and the snapshots would be the incrementals.

Question: But why does it causes further fragmentation?
Answer: Lets say that you have taken 7 snapshots.
(This means that you will have 7 incremental files.)
After that you take another one and you delete the previous 7.(This action is similar to the consolidation of the incrementals of an imaging application.)
The problem is that RX cannot really fill all the gaps of the free space that will be created between them, because some fragmented files have fragments in more than one of those 7 snapshots.
So you end with a new consolidated snapshot that will be fragmented to multiple pieces. (internal fragmentation of RX)
And here is the problem of RX. The internal fragmentation of RX+ normal fragmentation of the windows system, will make your system run like a turtle in the long run. And this is the reason that cannot be installed on servers.

The only way to circumvent this problem is once a month to:
a) Uninstall RX, defrag and reinstall RX.
or
B) Take a new baseline, use the defrag switch, defrag and take a new baseline again.

Panagiotis

[bgoodman4]

Thank you Peter and pandlouk

I think its slowly sinking in (and you are right, I do need to get my head around this interesting program). I can see why 3rd party defragers will not work and I can see why simply running Rxs defrager will not work either. I take it that when you establish a new baseline you are in effect deleting all previous snapshots (to keep with the normal imaging program description: all existing images and incrementals are deleted and a new image as of the current state is created). If I have this right then a usage strategy is beginning to emerge. Please correct me if I am wrong about anything.

Before establishing a new baseline use Rxs imaging tool to back up the drive and all its snapshots, uninstall Rx and defrag, or more logically establish a new baseline and defrag using Rxs defrag tool. Establish a new baseline, Image the drive again using the imaging tool.

This should (I hope) give me the ability to revert to an earlier state if necessary (some pre defrag state) or to the immediately post defraged state.

Is this a reasonable strategy and will it result in what I think/hope it will result in as described above?

[pandlouk]

Originally Posted by bgoodman4

Thank you Peter and pandlouk

I think its slowly sinking in (and you are right, I do need to get my head around this interesting program). I can see why 3rd party defragers will not work and I can see why simply running Rxs defrager will not work either. I take it that when you establish a new baseline you are in effect deleting all previous snapshots (to keep with the normal imaging program description: all existing images and incrementals are deleted and a new image as of the current state is created). If I have this right then a usage strategy is beginning to emerge. Please correct me if I am wrong about anything.

You got it right.

Originally Posted by bgoodman4

Before establishing a new baseline use Rxs imaging tool to back up the drive and all its snapshots, uninstall Rx and defrag, or more logically establish a new baseline and defrag using Rxs defrag tool. Establish a new baseline, Image the drive again using the imaging tool.

This should (I hope) give me the ability to revert to an earlier state if necessary (some pre defrag state) or to the immediately post defraged state.

Is this a reasonable strategy and will it result in what I think/hope it will result in as described above?

You cannot backup RX and all it's snapshots unless you choose a raw image and take the image outside the windows enviroment.

[bgoodman4]

I know what raw means in photography but not here, could you please clarify?

[pandlouk]

Originally Posted by bgoodman4

I know what raw means in photography but not here, could you please clarify?

Raw image = sector per sector image

[nexstar]

Originally Posted by pandlouk

You got it right.

You cannot backup RX and all it's snapshots unless you choose a raw image and take the image outside the windows enviroment.

There is one piece of software which can do this and that's Image For Windows using the Phylock utility. What it can't do is to get the MBR so you need a separate backup of that taken from outside of Windows. That only needs to be done once though after RB is installed.

Graham

[bgoodman4]

Originally Posted by pandlouk

Raw image = sector per sector image

Ah! thanks. So I can do this with True Image then and there would be no need to uninstall Rx before backup or re-installation of the image?

[Aaron Here]

Originally Posted by bgoodman4

Ah! thanks. So I can do this with True Image then and there would be no need to uninstall Rx before backup or re-installation of the image?

Yes, as long as you perform the sector-by-sector backup after booting up with the ATI boot disk.

[bgoodman4]

Thank you.