RESULTS of failed hard drive - SUCCESS!

[Azurik]

UPDATE June 28th

OnTrack recovered 30.6GB worth of data, and only 6 files were unrecoverable due to a missing/corrupted signature file. The exact cost to recover was $1,850.

Two lessons learnt:

#1. I will now BACK-UP BACK-UP BACK-UP!

#2. Research my data recovery service before choosing one. I could have saved $250.

UPDATE June 23rd

Hello AT Technicians,

I need further advise on this matter. As some of you know, I posted that my hard drive failed last week. After trying some things out you guys suggested without success, I went ahead and hired "Excalibur Data Recovery", based in MA, to try to extract the data for me. This is the following report I received from them. Any opinions on this? I am tempted to use the best company in the industry a try, OnTrack, to see if they can do anything about this - or am I just wasting my money for nothing (are all data recovery services the same?).

We are sorry to report that the diagnostics completed on the 40 GB Maxtor DiamondMax Plus 8 hard drive submitted were not positive. The technician assigned to this task indicated that initial standard bypass routines were not yielding positives results. The drive was believed to be suffering from electronic failure and a servo maintenance track problem.

The technicians? primary job was to stabilize the drive. A rebuild of the faulty electronics was completed. The first attempt was not successful at stabilizing the device. New parts were sourced and a second rebuild was implemented. This last effort was successful at steadying the drive to allow for a proper spin sequence. The technicians were then able to determine that critical tracks on the drive necessary to extract information cannot be accessed and the head armature can?t locate the servo track area.

Many attempts were made to gain access to the drive by circumventing the servo issue, however they were unsuccessful. The technicians then altered the proprietary software to fit the specs of the drive. After several alterations and tests were completed, the proprietary tools utilized were still unsuccessful in temporarily repairing the maintenance track. The drive must be able to identify the servo (which occurs prior to reading the boot record) before any file or directory accessing is possible. Since the storage properties of the device cannot be returned to manufacturer?s specifications, the data will not be recoverable.

The evaluation fee of $250 will be billed to the credit card number on file, which was previously supplied to Excalibur on the Data Recovery Agreement. We can return the media to you UPS Ground for an additional $15 (unless you request disposal). Thank you for your attention to this matter. We appreciate your business.

 

[JeffBruce]

I would call On-Track before puttng anymore money and tell them what Excaliber did,,,
jeff

 

[johnjkr1]

I think you should give drivesavers a call....fees to send it back..fee to look at it....remember my "nickel and dime...no data" comment?

 

[redbeard1]

When I was looking arouund, I received timely responces to my email inquiry to Ontrack. Make contact with both Ontrack and Drivesavers, eventually forward them the message, and see what they say. I'd say it doesn't sound good, but it won't hurt to ask.

 

[klaviernista]

Just so you know, what they are saying is that the hard drive head is unable to read the servo tracking signal of the disk. This is an unusual way for a hard disk to fail, but not unheard of. Most of what happens when you format a disk is the writing of a servo tracking signal (theres a lot more to it but essentially what it boils down to is writing a servo tracking pattern on the disk). This signal is what the head uses to actually locate date it has recorded on the disk. If your servo track goes bad, then the disk is basically useless because its basically looking for needle in a haystack.

The servo track can get messed up in a variety of ways. First and formost is exposure to a strong magnetic field (which generally ruins your recorded data as well). This exposure basically orients all of the domains on the hard disk in one direction, basically wiping everything (servo track and data).

Second, it could be a result of corrosion. The magnetic layer(s) of most hard disks is generally some type of cobalt based alloy. Cobalt is very prone to oxidation, so lots of technology has been developed to prevent this from happening, i.e. protective layers such as thin layers of diamond like carbon, silicon oxide etc.... If this protective layer gets damaged in anyway (can be microscopic damage) and becomes permeable to oxygen,the magentic layer can slowly oxidize. This can result in the disk losing its servo tracking pattern in some cases.

Third: It could and looks like it is in this case) be a reuslt of electronics failure. If the electronics controlling the write pole of the magentic head are faulty, the head could mistakenly overwrite portions of the servo tracking pattern when it writes data to the disk. I have no idea how you can fix this, which is basically what your HD repairmen are telling you.

I take it you had some pretty vasluable data on the 40 gig drive if you spent $250 bucks to try to get it repaired. It sucks, but chalk it up as a lesson to back up your valuable data periodically if you can't get it fixed. Good luck

 

[VirtualLarry]

Originally posted by: klaviernista
Just so you know, what they are saying is that the hard drive head is unable to read the servo tracking signal of the disk. This is an unusual way for a hard disk to fail, but not unheard of. Most of what happens when you format a disk is the writing of a servo tracking signal (theres a lot more to it but essentially what it boils down to is writing a servo tracking pattern on the disk). This signal is what the head uses to actually locate date it has recorded on the disk. If your servo track goes bad, then the disk is basically useless because its basically looking for needle in a haystack.

The only thing that strikes me as somewhat strange, is that I thought that most drives these days, used "embedded servo", not a seperate/dedicated servo surface or set of servo tracks.

Although, possibly, that's some sort of form-letter and/or oversimplification on the part of the data-recovery firm, trying to seem technical, but not get too technical. Then again, maybe they just screwed up and couldn't recover the data, and it was a convenient reason to supply for that failure.

Btw, to the OP, I missed the first half of this saga, can you possibly link to the original thread?

Originally posted by: klaviernista
The servo track can get messed up in a variety of ways. First and formost is exposure to a strong magnetic field (which generally ruins your recorded data as well). This exposure basically orients all of the domains on the hard disk in one direction, basically wiping everything (servo track and data).

Assuming that the platters are inside of the metal casing on the HD, then anything short of the magnet used to pick up entire cars at the junkyard, would not do that to a HD due to shielding effects. (AFAIK)

Originally posted by: klaviernista
Second, it could be a result of corrosion. The magnetic layer(s) of most hard disks is generally some type of cobalt based alloy. Cobalt is very prone to oxidation, so lots of technology has been developed to prevent this from happening, i.e. protective layers such as thin layers of diamond like carbon, silicon oxide etc.... If this protective layer gets damaged in anyway (can be microscopic damage) and becomes permeable to oxygen,the magentic layer can slowly oxidize. This can result in the disk losing its servo tracking pattern in some cases.

In that case, certain sectors could not be found, which in the case of the "hidden firmware sectors", might be necessary for the drive to properly initialize and load it's second-stage firmware. Then again, from what I understand, most drives have a redundant copy of that.

If the bad sectors aren't located in the "hidden area", then most of the accessable user data sectors should be accessable, assuming that the drive firmware doesn't lock up/spindown upon hitting too many prior sectors with unreadable servo info.

Originally posted by: klaviernista
Third: It could and looks like it is in this case) be a reuslt of electronics failure. If the electronics controlling the write pole of the magentic head are faulty, the head could mistakenly overwrite portions of the servo tracking pattern when it writes data to the disk. I have no idea how you can fix this, which is basically what your HD repairmen are telling you.

The infamous IBM 75GXP HDs apparently had a defect similar to this in their firmware. If power was removed to the drive during a write operation, the "write current" was not shut off, as the head retracted towards the spindle, thus wiping a nice spiral across all of the tracks, and blowing away the factory-written embedded-servo data in the process. Thus leading to the "clunk, clunk" of death upon the next spin-up attempt, because the heads couldn't locate the servo info. This was also why the W2K shutdown/APM/write-cache issue was so problematic. W2K did have a HD write-cache policy problem, but it shouldn't have caused outright device hardware failure - that was due to IBM's buggy drive firmware (apparently).

Originally posted by: klaviernista
I take it you had some pretty vasluable data on the 40 gig drive if you spent $250 bucks to try to get it repaired. It sucks, but chalk it up as a lesson to back up your valuable data periodically if you can't get it fixed. Good luck

Now I'm really curious what the back-story of this drive is. I love a good tech mystery.

 

[klaviernista]

The only thing that strikes me as somewhat strange, is that I thought that most drives these days, used "embedded servo", not a seperate/dedicated servo surface or set of servo tracks.

Yes and no. The method of servo pattenring used by a hard disk is generally pretty individual to the type of disk as I understand. Embedded servo tracking is certainly very attractive because it requires less magnetic material and is therefore cheaper. However, multilayer recording media using separate layers for servo tracking are not unheard of. Also, some HD manufacturers are exploring the use of physical servo patterning, where the servo pattern is literally an optically detectable pattern (similar to the data recorded on an optical audio cd) on the surface of the substrate.

Assuming that the platters are inside of the metal casing on the HD, then anything short of the magnet used to pick up entire cars at the junkyard, would not do that to a HD due to shielding effects. (AFAIK)

Again generally correct. The shielding provided by the metal casing will keep out most nasty external magnetic fields. The one it won't keep out is that generated by the magnetic head which is in the enclosure itself. If the head controller is faulty, the head could be generating a field inside the casing that would alter the servo track in unpredictable ways. As heads are specifially designed to apply a magnetic field having sufficient intensity to overcome the coercive force of the domains of the recording media, this is a legitimate concern.

 

[Azurik]

UPDATE June 28th

OnTrack recovered 30.6GB worth of data, and only 6 files were unrecoverable due to a missing/corrupted signature file. The exact cost to recover was $1,850.

Two lessons learnt:

#1. I will now BACK-UP BACK-UP BACK-UP!

#2. Research my data recovery service before choosing one. I could have saved $250.