Does Rescue and Recovery have any utility to fix bad sectors

[bonez318ti]

I dropped my laptop a few days ago and I can't boot into windows to run chkdsk. It says HAL.DLL is missing, when I run the diagnostics, it shows that there are a bunch of bad sectors. I need to recover my data (that I havent backed up in 6 months) before I format, or perhaps get a new drive..

so my question is.. I understand that chkdsk can recover data from bad sectors, is there a way for me to fix these sectors using IBM's preloaded R&R?

Thanks

[Moderator Edit: As the ThinkPad is an X41, the message thread has been moved the X series forum]

[EOMtp]

If you have a real problem, then nothing -- absolutely nothing -- recovers data from bad sectors like SpinRite:
http://www.grc.com/sr/spinrite.htm

Chkdsk will destroy your data. Chkdsk is fine for removing lost clusters, but don't even think of using it to recover data.

Note: You may need a floppy drive to boot SpinRite if your hard disk is not set up to boot to an MS-DOS Command prompt. SpinRite will NOT run under Windows because what it does so well is impossible to do under Windows constraints.
Good luck!

[DIGITALgimpus]

There's no real way to "fix" them. Only mark them so your disk avoids them.

AFAIK SpinRite pretty much has the best reputation for dealing with them.

Overall it's risky regardless, since a bad sector is a bad sector regardless of what you use.

[EOMtp]

You say: "There's no real way to "fix" them. Only mark them so your disk avoids them."

That is as true literally as it is besides the point to this discussion. The issue is NOT that of fixing bad sectors, but rather RECOVERING the data in them ... and the latter is most definitely possible, albeit with varying degrees of success.

Further, there is nothing "risky" about making the attempt to recover the data, i.e., giving it your best shot.

I mention all this because I do NOT want the person who started this thread to think that he might as well give up because there is low probability of success or some risk in undertaking the venture. Neither is true.

[christopher_wolf]

It is indeed a Risky Business trying to fix anything that went "bad" on a HDD. The best that SpinRite does, as it has a good reputation for this, is to re-map the sectors so that the drive doesn't use it again, this is something the OS has to be aware of as well, and recover the data that happened to be stored there, which is more interpolation and educated guesssing on its part based on the quality and contiguity of the related data structure that just so happened to include the bad sector.

[EOMtp]

With respect, there seems to be a misunderstanding of what "risk" means. There is NOTHING risky about trying to recover data. People seem to be confusing "risky" with "not guaranteed to succeed".

[christopher_wolf]

With respect, there seems to be a misunderstanding of what "risk" means. There is NOTHING risky about trying to recover data.
People seem to be confusing "risky" with "not guaranteed to succeed".

No no, not at all.

I believe it acquired the latitude of being used as a misnomer in this case. As with any data recovery from a HDD with issues, it works best in an ideal environment, such as sparsely distributed bad sectors with a good deal of related information that allows it to successfully recover/re-interpolate the part of the data that assigned to the bad sector. If you have an old HDD riddled with bad sectors, then the job becomes much more difficult accordingly.

[bill bolton]

There is NOTHING risky about trying to recover data.

There are very real risks, not the least of which is being able to determine whether any particular piece of data that is technically "recovered" accurately represents the data that was originally written to the disk sector concerned.

Cheers,

Bill

[EOMtp]

This conversation is most fascinating. One would not believe that there would be this much confusion between the meaning of "risk" in undertaking a task and the expectation of success in that undertaking. One more time: In the context of the issue at hand, "not guaranteed to succeed" is not synonymous with "taking a risk" or exposing oneself to "very real risks".

As an example of REAL "risk", consider the following: Boosting the power applied to the disk drive in an attempt to determine if the heads become more sensitive to reading the data off the bad sectors.

[bill bolton]

One more time

You can say it as often as you like but it doesn't make it correct.

Cheers,

Bill

[EOMtp]

"Risk" means "statistical odds of DANGER". What dangers do you foresee in attempting to recover data? We should not be arguing over the simple meaning of simple English words ...

[christopher_wolf]

"Risk" means "statistical odds of DANGER". What dangers do you foresee in attempting to recover data? We should not be arguing over the simple meaning of simple English words ...

Risk can have many meanings (a quick link; See, http://webster.com/dictionary/Risk ), ever taken an advanced statistics course? The best job that it can do is interpolate what data was there in the sector, based on the quality and relationship it had with the rest of that data structure. It doesn't "automagically" regenerate material out of nothing; that simply doesn't happen. Statistical odds of Danger? Whoa, there....Not quite.

For it to be true, statistical odds, you have to repeat it and come up with a good, arbitrarily close, value for the odds of it giving you either a perfectly function restore, a non-functional/garbled restore, and all the states in between on the same drive under the same conditions, with the same data. You, I believe, are referring to the colliquial/casual interpretation of the word "Risk." To be really precise, there is no "chance" of a change in the weather tomorrow, that is just an educated assumption based off of a set of, very complex, conditional probabilities. Same with "Risk" and the recovery of data; it is a conditional probability. Note this agrees and is consistent with the way it "recovers" the data by interpolating what must have been there from its absence and the, assumed, interaction it had with other members in the same data structure.

There is *always* a risk associated with doing that, no matter how small. It is essentially taking very educated guesses as to what the content was...more of a deductive process; yet, as any deductive or inductive process, it is only as good as the data and the interpretation thereof

Yes, I have seen recovery software messup a file recovery from a HDD/storage device. like that; sometimes by very little, sometimes alot. That possibility always exists, although it may be very small.

[EOMtp]

Christopher: What you state about statistical analysis is all true, but not pertinent to this discussion. You are assuming the the wrong model of what recovering data off bad sectors entails from an engineering perspective.

The process is not one of interpolating data. The process has to do with repeated attempts -- under different initial conditions and with more repetitions than are embodied in any standard operating system algorithms -- at reading a sector in question until at least one attempt succeeds. Success, if/when achieved, is unambiguous because of the Hamming codes which are used to guarantee the accuracy of data transfer.

[bill bolton]

until at least one attempt succeeds. Success, if/when achieved, is unambiguous because of the Hamming codes which are used to guarantee the accuracy of data transfer.

That is not a reliable measure of successful recovery, even for normal business purposes. It is regarded as legally unreliable for forensic evidenciary purposes.

Cheers,

Bill

[EOMtp]

Bill, we are talking about some guy trying to decide if his collection of MP3s is still there -- not whether we should stake the future of mankind on the accuracy of these data.

[christopher_wolf]

No, but there is still the possibility...Nothing big, I know; but there isn't exactly no risk either. For practical purposes? There is a pretty high conditional probability that it will go OK. Nobody said it would fail or there is a "significant risk to the data;" there just isn't no risk. Negligible risk, yes, but it is highly unlikely that anything will go wrong in a practical, real world setting.

There is a quite a good number of recovery algorithms; the one that you mentioned simply re-reads the sector until it gets a meaningful result, a very simple way. Plus, the more passes you make, the better it gets....Let it go for awhile and you will recoer a healthy portion of data. Others, such as the one I was referring to, would perform an interpolation and restructure the data entries. The advantages of that are that it is faster and, given enough time and input on part of the operator, more accurate and it requires less operational/mechanical strain on the HDD (which would increase the probabilities of a headcrash or other failure event, making things even more difficult); disadvantages include a higher rate of mistakes if the data used wasn't filtered/screened carefully and implausible boundary conditions are applied. There are further algorithms that assume a certain amount of redundancy, only useful if there actually is any to begin with, and take it from there.

I have also done my fair share of managing a lab, Biotransducers at the time, and as a grader for various math, science, and engineering science courses (statistics, abstract algebra, MSE, etc). I have a good 4 years of Cory Hall, Soda Hall, and EECS under my belt. Great fun it is, isn't it?

[EOMtp]

Yes, Christopher; great fun, indeed! Glad you are enjoying it.

[bill bolton]

some guy trying to decide if his collection of MP3s is still there

Hmmmm..... my bad. I should have realised much sooner that were trolling.

[EOMtp]

[quote="bill bolton"]Hmmmm..... my bad. I should have realised much sooner that were trolling. :shock: :([/quote]

Bill, No reason to follow ignorance with insult. The point -- which you may have missed -- was that one should not lose perspective of the task at hand, i.e., not to let the best become the enemy of the good. Let's end this as it's no longer of technical value to the person we were trying to help. Have the last word. Best regards.

[bonez318ti]

To follow up on this thread, I used downloaded knoppix and used that to copy my data (most of it was in a separate partition) and I beleive most of it is safe. (I didn't bother with my MP3s )

Now, the question is, which folders/files can I copy over from the C drive so that I''ll be able to quickly reset all of my preferences and stuff when I get my new hard drive (which is SOLD OUT EVERYWHERE!!!!.. well atleast Zipzoomfly and Newegg... any other places to get it??)

Thanks

[EOMtp]

Much (most) of that configuration information, including application-specific information, resides in the Registry. No easy way -- and in some cases no way -- to "copy" it over to a fresh Windows XP installation.

Your best option is to try to "reconstruct" Windows XP on top of the existing damaged installation. That's the ONLY way you will achieve precisely what you want.

You will need the facilities (equipment and CDs) to do the following:
1) Clone (as much as can be recovered) from the old disk to the new. If necessary, copy every file "the hard way".

Note: After you transfer as many files as you can, you may wish to try the steps which follow on the original (i.e., presumed damaged) disk drive. If that does not work, then proceed with the new drive onto which you copied the files.

2) Boot from an XP CD and see if it finds an "existing" Windows XP installation on the disk.
3) If yes, then install XP on top of the existing installation and you will have what you want, i.e., everything as it was.

If an existing XP installation is not found on the target disk, then I regret to inform you that you are looking at installing XP from scratch and building/installing your applications anew.

By the way: What is the make and model of the damaged drive?

[bonez318ti]

I am using an X40 that originally came with a 40 gig hitachi harddrive. I beleive it is the 1.8 form factor and its the hitachi travelstar C4KX0 line.

I tried earlier to pop a win xp cd in there and i dont think it had the option to reinstall over an existing installation.. so I guess I'm out of luck.

Is it possible to drag the documents and settings over from my user profile? I want to transfer over things like my bookmarks in firefox/IE.

[EOMtp]

The option to install on top of an existing XP installation appears late in the game. First you are asked if you want to recover from recovery media -- answer NO. Then you are asked if you agree to the license and if you want to install XP -- answer YES. Then the installation program searches your hard drive for previous XP installations and reports what it found. This is when it asks if you want to install in the same folder as a previous installation it found -- answer YES to this question if the previous installation was found -- or install in a new location. Did you do this and did it find a previous XP installation on the disk?

Yes, it is possible to transfer profile folders from the Documents and Settings folder over to the new system. Transfer your user profile PLUS the "All Users" profile -- and many settings, not all, will transfer to the new system.