Frag me

Background info on how Windows defrag works for Vista http://blogs.technet.com/b/filecab/archive/2006/07/10/disk-defragmenter-faq.aspx

and also Win 7 improvements http://blogs.msdn.com/b/e7/archive/...d-engineering-the-windows-7-improvements.aspx

 

Ok folks, for the third or fourth (and last because this is getting too frustrating) time, Comodo Time Machine and ALL other products that were developed from Rollback RX, CANNOT WORK WITH A DEFRAGGER, not even the good, ole, built-in Windows defragger. AAAAAAHHHHHHH!!!!!!!

Acadia

 

@Acadia: A separate discussion developed within this thread, regarding the efficacy of third-party defrag/optimization tools in general. Inadvertent hijack, sorry.

@Bill_Bright: If you intend to reply to my latest, perhaps it would be best to start a new thread. Unless the response specifically deals with the CTM situation.

 

CTM is based on Rollback RX - which uses its own block level redirection technology using a driver filter. When changes to a block are made, it marks them as protected and redirects changes to new blocks. As this is at a block level this also means the defrag tools cannot defrag any blocks which have been marked protected.

CTM is compatible with defrag tools as far as it won't cause data corruption, but it will/can reduce the effectiveness of defrag tools.

To compare to VSS (used by built in Windows system restore and other tools), VSS does not protect the blocks on disk from being moved, which means if the defrag is not VSS aware it can move the shadow data, which can invalidate existing snapshots by making a current/new snapshot huge and causing old ones to be removed.

To make a defrag tool VSS safe, all you need to do is move data in 16KB chunks (this appears to be the granularity that VSS uses and I don't know why it was chosen) and the Windows defrag API will ensure that the volume shadow space does not increase (and if it gets too big pushes out older shadow copies from your drive)

The only way to prevent fragmentation building up with CTM/Rollback RX would be to regularly defrag in the hope you defrag files before they get modified and the sectors they reside on get marked as protected.

 

Agree with Bill.
Mrk

 

Windows prefetcher is actually part of the memory manager and stores traces in the prefetch folder, the memory manager uses these traces to optimise loading into memory executable and data from disk when the process in question is called.

As a related but separate process a layout.ini is created from these traces. This is used by the built in defrag tool of windows (and any defrag tool) to optimise the placement of files required for prefetch purposes.

 

They go hand in hand but I was mainly talking about the defrag process.

But to "operations" first, remember, there are two critical specifications we are talking about here. Seek and read times. Reading the contents of any given single sector is essentially the same regardless the level of fragmentation or location on disk. "Seeking" those locations is the spec that counts here - that is, the time it takes for the R/W head to physically move across the platters, seek out and position itself over the "first" file segment in that file. "First" is the critical word there.

After defragging, once the beginning ("first" segment) of the file has been found, the rest of the file segments are all right in line. Therefore, the performance and "efficiency of operations" (at least in terms of opening the files) is the same.

HOWEVER because Windows 7 and especially Windows 8 use their own defragger and fetch routines to "optimize" file locations for faster opening based on the users computing habits, if you consider faster Windows and program load times as more efficient then using Windows defragger does result in a more efficient "operations".

As far "efficiency of the defrag process" what I mean there is even if Auslogics achieves 100% (just tossing arbitrary numbers out) defrag efficiency (that is, every single file on the disk is saved with each file segment adjacent to the next segment), and Windows defrag only achieves 97%, I am saying that (1) 3% is negligible to begin with and most importantly (2) that 3% advantage quickly levels out as soon as you start using the computer.

But even those 3% files will be mostly defragged into a few large fragments instead of many smaller fragments. So the R/W head has to jump .5 inches 3 or 4 times instead of jumping to the adjacent segment. Even with today's 5400RPM drives, we are talking just a few (like single-digits) milliseconds!

Again, YEARS AGO with that old antiquated XP, slow and tiny buffered hard drives, and small amounts of RAM, getting the most efficient "defragging" of the files on our disks justified third party defraggers.

But with today's much faster drives, huge amounts of RAM (= less dependence on page file), the vastly improved integrated defragger and the "optimize based on user's usage" features makes the use of third-party defragger a bad idea - or at least of waste of disk space, and time.

Finally, I emphasize, if you need regular defragging, you don't have enough free disk space.

 

@Acadia - and for the umpteenth time - got a link?

@Windbringeth - not sure why I am singled out here. I note this is the 32nd reply to a thread the OP has yet to return to after his opening post.

 

First, I sincerely apologize for chiming in again when I said that I would not, but this was too temping to resist: sometimes it is necessary to cut off one's own finger in order to save one's arm:

http://horizondatasys.com/files/downloads/RollBack_Rx_User_Guide.pdf

page 40 at the very bottom

(Again, PLEASE realize that Rollback and CTM are the same thing!)

Acadia

 

Thank you. Please note what it says is quite different from your claims.

Not true. They will work. It just does not make sense to use them.

Again, not true. They can work - again, just not needed as CTM contains defragging features and you should only use one defragger, regardless which one.

Now this was not just incorrect, but totally wrong. "Royally messes with your sectors" suggests drive corruption! Damage to our system and otherwise destructive behavior. That is totally false and does us all a disservice.

So I would have to say all the drama with "third and forth time", your frustration and AAAAAAHHHHHHH!!!!!! were a bit over the top - considering it was all inaccurate.

 

Ok, people, I am TRULY pulling out of this thread, the current logic confounds me. In the past I have usually considered myself somewhere in the middle of understanding pc science but this particular case has me going in circles ...


Acadia

 

Cool,I was not aware that it catches and prevents all block level access.

 

They will work only for the non protected partitions/drives. For the protected space they won't because the defragmentation apis will be intercepted and discarded before reaching the hdd firmware.

You do not know what you are talking about.
CTM defrag feature is not a defrag, only reclaims/frees the disk space that ctm had marked as read only but does not contain any valuable data anymore.

Acadia expression was(?) wrong; he did not mean physical damage of the sectors but corruption of the data on those sectors; The data on the sectors can be corrupted when CTM defrags (or updates the baseline or gets uninstalled); something expected when over a file system (NTFS or Fat32) you place another hidden file system (CTM's) that treats the main file system as a single locked file/container/virtual HD, and at the same time excludes some files like the hybernation and the page files from that container.

Bill if you never used, for not talking about investigating, how CTM works, where it places its hidden file table, how it treats the real (baseline) file table, each snapshot's (virtual/hidden) file tables and especially how it achieves the system protection by marking as used (blocking write access to) every cluster that gets written until it's own defrag takes place and frees those clusters for the OS to use, you should refrain from accusing others of being inaccurate.

Panagiotis

 

"I" don't? Yeah, right.

Note the last sentence on page 40 of the user guide link provided by Acadia,

Where did I say "physical" damage? I didn't. I said, "drive corruption" and you quoted it.

Then you first say, "CTM defrag feature is not a defrag" then immediately you say, "The data on the sectors can be corrupted when CTM defrags". Make up your mind.

But that said, IF you know CTM corrupts data, why use it?

 

Good grief, I feel like Michael Corleone in the third GodFather: Just went you think your are finally out they drag you back in (I wasn't referring to you, pandlouk).

I have NEVER, EVER used CTM, not even in trial. I feel safe in saying that pandlouk does not use it either, he can correct me if I am wrong.

I was just trying to help everyone understand a program that I definitely do NOT recommend and have ALWAYS directed folks away from!

Acadia

 

The fact that they named it "defragmenter" means nothing to me.
"Drive", last time I checked, in hardware was a physical device. So "drive corruption" is physical. Data corruption is another thing (for sure not "drive corruption") and in case you meant data corruption when you responded at Acadia well... you were totally wrong... and not Acadia.

I know pretty well, what I write, I just used the same terminology in the same wrong way that the manual and the program uses...

And who told you that I use it?

Panagiotis

 

Sorry Rico. Your topic was taken WAY off course. My apologies for my part in that.

I hope you were not driven off.

 

Rico, try mydefrag, but remember to check FAQ beforehand:
http://www.mydefrag.com/Manual-FrequentlyAskedQuestions.html

Meanwhile this topic has too many OT posts.

 

I've just finished uninstalling Comodo Time Machine and doing a defrag with Auslogics Pro and the difference in disk performance is incredible! I had CTM Installed for a long time so defrag has been impossible for some time.The uninstall went flawlessly "Thank God".

I would really hope Comodo finds a way around this compatability issue with doing a defrag while it's installed because its a real pain and time consuming.