HTS543232L9A300 not spinning

[XXL]

2.5", 320G, SATA
0A57128
DA2358_
Z1P833
0MGD

The motor does not spin and therefore the drive isn't detected on POST (no delay in detection time either). The PCB visually looks fine. When powered on - the drive itself is making a hardly audible low-pitched noise (like a beep) within an interval of 3 seconds or so. This has happened after air transportation (traveling by plane) - prior to that the drive was fine. What could be the probable issue?

Thanks.

[hddguy]

Without professional diagnosis it is difficult to identify the exact problem.

I would assume that the spindle is trying to power but badly parked heads are causing resistance and the audible sounds you hear is failed attempts at starting the motor.

If this is the case each time the media is powered the attempts to start the spindle while recording heads are 'stuck' to the surface is causing more damage.

You need to get this professionally diagnosed. Further attempts to power or DIY repairs will probably only result in further complications.

[poehere]

We have a lot of respected DR techs in Europe but need to narrow down your location to say which one is closest to you. You will have to have this looked at.

[XXL]

Without professional diagnosis it is difficult to identify the exact problem.

I would assume that the spindle is trying to power but badly parked heads are causing resistance and the audible sounds you hear is failed attempts at starting the motor.

If this is the case each time the media is powered the attempts to start the spindle while recording heads are 'stuck' to the surface is causing more damage.

You need to get this professionally diagnosed. Further attempts to power or DIY repairs will probably only result in further complications.

Thank you for the reply. If you don't mind I would like to ask some additional questions.

So, could the MHS possibly been misaligned from the parking zone due to physical impact during transportation? And also, what would your next step be as a data recovery specialist if you had to deal with this HDD - is it taking the cover off to inspect the position of the heads in a clean room environment?

[northwind]

This drive is most probably suffering from stiction (at least).
This means that you can take the safe route and consult a pro, or try diy and complicate things even further.

I don't understand why you ask the additional questions. If you want to attempt DIY by opening up the cover etc using someone's guidance then i'm afraid no pro would consult you or guide you to it.

[pcimage]

We can help in UK, and if it is stuck heads then it shouldnt be too expensive.

But if you do open the drive and try DIY, expect price to be 2x or 3x at least and much less chance of recovery.

Good luck

[XXL]

This drive is most probably suffering from stiction (at least).
This means that you can take the safe route and consult a pro, or try diy and complicate things even further.

I don't understand why you ask the additional questions. If you want to attempt DIY by opening up the cover etc using someone's guidance then i'm afraid no pro would consult you or guide you to it.

Before taking it to a tech. I would like to know as much as possible about the potential procedure that is going to be undertaken and learn a few things here and there while I have the chance - is that a bad thing? I am fully aware of the DIY risks, dangers and misconceptions and I will not attempt any DIY on the drive in question - I will indeed hand this HDD over to a DR specialist. However, as I have a lot of "broken" drives at my disposal as well - I would like to get myself familiar with at least some occurring issues.

Today, I decided to pop the cover off a 2.5" WDC drive that had a similar problem - spindle wasn't working, long 2 sec beep, the system was stalling on POST while trying to detect the drive on the connected channel (difference to the Hitachi above). This was purely for educational purposes and minor experimentation - this was going nowhere to start with as I have contaminated the drive the minute I opened its cover in a dirty environment. What I witnessed was "stuck" heads - I tried moving the arm away (without lifting heads) with my toenail grabbing it by the side and it worked after 2-3 tries. As I understand - this is a horrible method as the heads could leave scratches on the surface if they are not lifted? The arm ended up on the parking ramp in the end. Just for kicks I assembled the drive (I just blew air on the platter surface to remove easily-movable dust particles - pretty sure I managed to move only a fraction while smaller ones remained).

To my shock and surprise the drive survived this procedure. The drive spun up and was properly detected on POST without any signs of trouble. Afterwards I was able to mount both of the NTFS partitions on a linux distro. However, due to having dust over the surface, I assume that I will now get heaps of CRC errors over the stored files and bad sectors. So the data must be clearly toast to some extent.

I understand that the actions undertaken above bear no judgement. However, I would possibly like some input on the whole thing - basically comments on how are the things above done properly (tools, technique), if I may ask that?

Also, are there any known ways to clean the platter surface? And will a drive with "stuck" heads NOT halt/stall on POST and the system continues booting as if nothing connected to that channel (nothing detected as connected on POST)?

Thank you for any input. And as I have already mentioned - I will definitely not mess around with the Hitachi as the data stored there is valuable, so I will be taking it to a DR tech. I am located in Estonia.

[hdrescue]

Hello

That is badly parked heads almost for sure. And like the guys told you, you have really good DR professional in Europe.
I work from Portugal and I can help you to if you need.

You know the drive need to be open in a clean environment and more than that is about knowledge and DR specific tools.
Of course we don´t share information that takes years to learn, develop and investigate.

[HDD Spaz]

Just when I think I have seen it all - Toenail recovery!

[XXL]

Just when I think I have seen it all - Toenail recovery!

LOL. Sorry about that. What I meant was "fingernail". No idea why I stated "toe", guess it's due to my over-broken english practices. Actually, laughing quite hard about this right now.. Imagining the picture how i'd hold my foot next to the drive to "unstick" the heads with one of my toes

That's the side area marked on the img. where I applied pressure with my fingernail in order to push the arm to the ramp.

Attachments

stuck_heads.png (139.39 KiB) Viewed 12213 times

[XXL]

Hello

That is badly parked heads almost for sure. And like the guys told you, you have really good DR professional in Europe.
I work from Portugal and I can help you to if you need.

You know the drive need to be open in a clean environment and more than that is about knowledge and DR specific tools.
Of course we don´t share information that takes years to learn, develop and investigate.

Hi. Thank you for another confirming point of view.

I've already got in touch with the local DR office today - will see where that brings me.

Sorry, for the latter part of your message I have to express a bit of dissent and find it unfitting in the corresponding context. Especially because it has little to do with the matter that I am asking about. All I asked was some trivial tips or bits and pieces that this forum was full of - this information has already been freely shared and discussed many times, it's just that its all over the place and not accumulated. So, I am not sure why are you using such misleading wording (as if I asked to teach me firmware or something as extreme). Although I have noticed that there are 2 types of people on this forum: materialistically motivated highly-protective conservatives and people that are understanding enough to share the knowledge/experience that they have acquired while being in the same position as the people asking for help in their time. However, I am sure this is an old argument, no point in going these lengths.

[hddguy]

Here is an example of stuck heads:

http://www.youtube.com/watch?v=7c_MTX6-gYg

In this example the slider has become detached from the head assembly. Attempts to repair by the 'technician' here results in further damage and parts would then be required, though the chances of damage to the media surface would now be significantly higher.

It is very possible the magnetic heads in your disk are also damaged, and possibly your case cannot be restored without the use of parts, but all this needs to be identified during a professional diagnosis.

Once the HDD is operational it will very likely have some logical damage to the surface, i.e bad sectors. Even for this specific recovery or cloning hardware / software is required. Even at this stage your data would not yet be safe.

there are too many potential complications for DIY.

You are right in that there is lots of information about this problem here, and the fact that there is a 'Search' button means it can be found with minimum effort on your part.

Mostly, people do not have the time to explain in detail the same problem over and over each time a new member asks about it.

People here are helpful and quite a lot of information is shared, though in my opinion this only encourages DIY which results in increased costs with a professional lab and compromises the recoverability and safety of potentially important data.

What you decide is your choice, but realistically you need to consult a professional who can handle this

[northwind]

Before taking it to a tech. I would like to know as much as possible about the potential procedure that is going to be undertaken and learn a few things here and there while I have the chance - is that a bad thing?

Not at all. Following instructions you read on forums/on youtube/in a blog/somewhere written by someone somehow IS bad.
Y' know, it is very easy -even for a very good tech- to make mistakes, esp. if that someone does not have the drive in front of him. So lets say, you take advice by someone who tells you, hey! it's easy, all you have to do is hit the drive with a hammer (don't laugh my friend, it is a REAL example, just search THIS forum). And you do it. And then you find out that what you did had no good result and you decide to take your drive to a pro who opens up the drive and sees this:

Know where i'm getting at?

Anyway, no matter as i read you have decided to take the safe route, and hddguy has given very good explanations to your questions, so good luck

PS. I fell off the chair laughing with that toenail. Funny thing is, i didn't notice it until HDD Spaz mentioned it

[XXL]

Here is an example of stuck heads:

http://www.youtube.com/watch?v=7c_MTX6-gYg

In this example the slider has become detached from the head assembly. Attempts to repair by the 'technician' here results in further damage and parts would then be required, though the chances of damage to the media surface would now be significantly higher.

It is very possible the magnetic heads in your disk are also damaged, and possibly your case cannot be restored without the use of parts, but all this needs to be identified during a professional diagnosis.

Once the HDD is operational it will very likely have some logical damage to the surface, i.e bad sectors. Even for this specific recovery or cloning hardware / software is required. Even at this stage your data would not yet be safe.

there are too many potential complications for DIY.

You are right in that there is lots of information about this problem here, and the fact that there is a 'Search' button means it can be found with minimum effort on your part.

Mostly, people do not have the time to explain in detail the same problem over and over each time a new member asks about it.

People here are helpful and quite a lot of information is shared, though in my opinion this only encourages DIY which results in increased costs with a professional lab and compromises the recoverability and safety of potentially important data.

What you decide is your choice, but realistically you need to consult a professional who can handle this

Thank you kindly for the reply. I have already seen the referenced video some time ago and that indeed serves as a good alert in regards to the complexity level of the performed operation.

As I have already stated - I will hand this drive over to a DR technician. However, I would like to clarify a few points that I can learn from. If the magnetic heads are damaged - how is this usually diagnosed in the lab? What sort of tools are used (microscope?) / hardware (PC3k?)? So, if the head stack is damaged - assuming I would need to swap the heads - could you name a few popular tools for that? This will enable me to do more thorough research. Of course I understand that these sort of things are like a broken record over and over - I do use the search feature as much as I can - it's just I'm not able to get the keywords right all the time for relevancy.

My question still stands - how does one clean the surface of the platters? Supposedly canned air can be applied gently, but what about particles that are hard to remove with this method? Is there any applicable/known working cleaning liquid/substance?

Regarding the logical recovery - I have had some modest (about 6-8 drives) experience with cloning drives with pending bad sectors in gddrescue and dd (iflag=direct) and doing the remap process manually by writing NUL data into the troubled sectors / repairing the fs afterwards. Now that I have the chance, I woud like to ask why is gddrescue perceived to be a better tool than dd - they seem to recover the same amount of data? And also, regarding hdparm, it seems that no matter what I do - it always fails to put the drive into a PIO mode reporting a sg_io error, bad sense etc (I'm probably missing something)?

I can clearly see the obvious disadvantages due to attempted DIY as a onetime stand and the havoc it might cause, but I guess there is no way to change human nature - people will keep making these mistakes no matter what. It's a matter of using the acquired information - either intelligently or like there's no tomorrow.

Thanks a lot.

[Vulcan]

I woud like to ask why is gddrescue perceived to be a better tool than dd - they seem to recover the same amount of data?

In theory it would be possible to duplicate the behaviour of GNU ddrescue (package gddrescue) by using dd, if the technician involved was very skilled and if they made no mistakes during the process. However in practice, using ddrescue requires less skill (though it still needs some skill for best results) due to its built-in algorithms & logging, and typically has better & less-risky results, partly due to preventing the human errors which could otherwise easily occur if trying to use dd manually with many different parameters, and making it easy to control the relevant tuning of those parameters. The specific benefits of using ddrescue also depend on the problems with a given disk (e.g. just one unreadable block on the disk, or 1000 marginal blocks in different places on the disk which might be readable with extra retries, or ... etc. etc.)

In short, the advantages of using ddrescue make it a "no-brainer" choice over dd, for reading from imperfect disks (and in some other situations), in my experience.

(Note: This is just comparing ddrescue & dd - it does not address the risks / problems which are common to using either program.)

[XXL]

I woud like to ask why is gddrescue perceived to be a better tool than dd - they seem to recover the same amount of data?

In theory it would be possible to duplicate the behaviour of GNU ddrescue (package gddrescue) by using dd, if the technician involved was very skilled and if they made no mistakes during the process. However in practice, using ddrescue requires less skill (though it still needs some skill for best results) due to its built-in algorithms & logging, and typically has better & less-risky results, partly due to preventing the human errors which could otherwise easily occur if trying to use dd manually with many different parameters, and making it easy to control the relevant tuning of those parameters. The specific benefits of using ddrescue also depend on the problems with a given disk (e.g. just one unreadable block on the disk, or 1000 marginal blocks in different places on the disk which might be readable with extra retries, or ... etc. etc.)

In short, the advantages of using ddrescue make it a "no-brainer" choice over dd, for reading from imperfect disks (and in some other situations), in my experience.

(Note: This is just comparing ddrescue & dd - it does not address the risks / problems which are common to using either program.)

I see. Could you elaborate a bit further on gddrescue, namely are there any other useful built-in algorithms apart from a block-size-changing one? I found the benefits of this feature to be somewhat of little help when dealing with a disk with large amount of bad sectors due to extremely long execution time - gddrescue will only decrease the block size on the second run, which will trigger the same delay when encountering a bad sector. So, the execution time is already at least doubled. Won't the best practice in this case involve a simple run of dd with the lowest block size (512B), instead? This is, of course, in regards to a drive with many bad sectors - AFAIU, running gddrescue with a large block size on a drive with only a few bad sectors will be more desirable speed wise, as thanks to logging gddrescue can start an automated N number of passes, while adjusting the block size, for attempted post-recovery.

I've mostly ran dd with only 3 flags: direct, sync (for padding with NUL) and noerror, all of this with a block size equal to a sector - this has given me the best results (although noticeably slow due to direct I/O). Is there any other flag that would influence recovery for dd? Or, perhaps, some feature of gddrescue that is able to do even more?

I have to ask again - how can I put the drive into PIO mode (from what I read this should aid recovery)? Presumably this is done through hdparm -X? This command never worked for me, for some reason. The error I'm getting is something like this: SG_IO: bad/missing sense data. What other known working method is there for PIO mode switching?

At the very end of your post you mention the risks associated with running these apps - may I ask what are they and what is perceived a better alternative (somewhere I read about Data Extractor feature of PC3K)?

Finally, when making a raw binary image with dd or gddrescue, is there a way to somehow evade the massive delay (I/O err of about 10-15 seconds) when encountering a bad sector? Perhaps, some kernel setting? A surface scan with mhdd goes significantly faster than the time spent upon detection by the Linux kernel - I guess, this is influenced by the error reporting mechanisms or large predefined timeout values in it for such encounters?

As of again, any input on this is greatly appreciated!

[Vulcan]

I counted around 14 separate points / questions where you seem to be expecting a reply - some of which would need multiple paragraphs per response. At a conservative 10 mins per detailed paragraph, that's over 2 hours of my time that it would take to reply, so I'll politely decline to do that. Perhaps someone else here is less busy and has that time... But here are a few brief comments.

Some of your questions are answered in the online ddrescue manual (e.g. the algorithm section - changing the block size is not all that it does); several others are hypothetical questions where the data necessary to give a detailed reply are not (and cannot be) provided by you, without having a specific faulty disk available on your system and mine, to test any hypothesis. So for many questions, the answer is "it depends on the specific behaviour of the disk...". Faulty disks can behave in many different ways which are different to your suggestions above. I've learned not to get into such hypothetical discussions here - they just go on & on, as it is impossible to know when a "correct" (or "the best possible") answer has been given.

Finally - yes, you should expect different behaviour between dd / ddrescue & MHDD, because the error recovery / retry / timeout behaviour in Linux drivers are completely different to MHDD. That does not mean that ddrescue isn't a useful tool, when used in ways that minimise its limitations, in an amateur (i.e. not a professional) recovery situation, if nothing better is available to perform cloning, and if the risks of such DIY are accepted. Modifying the OS drivers to adapt their behaviour can also help, but that's a whole different skillset...

I hope someone with more time comes along and answers all your questions, but personally I think you might be expecting a bit too much with your level & quantity of questioning.

[XXL]

I counted around 14 separate points / questions where you seem to be expecting a reply - some of which would need multiple paragraphs per response. At a conservative 10 mins per detailed paragraph, that's over 2 hours of my time that it would take to reply, so I'll politely decline to do that. Perhaps someone else here is less busy and has that time... But here are a few brief comments.

Some of your questions are answered in the online ddrescue manual (e.g. the algorithm section - changing the block size is not all that it does); several others are hypothetical questions where the data necessary to give a detailed reply are not (and cannot be) provided by you, without having a specific faulty disk available on your system and mine, to test any hypothesis. So for many questions, the answer is "it depends on the specific behaviour of the disk...". Faulty disks can behave in many different ways which are different to your suggestions above. I've learned not to get into such hypothetical discussions here - they just go on & on, as it is impossible to know when a "correct" (or "the best possible") answer has been given.

Finally - yes, you should expect different behaviour between dd / ddrescue & MHDD, because the error recovery / retry / timeout behaviour in Linux drivers are completely different to MHDD. That does not mean that ddrescue isn't a useful tool, when used in ways that minimise its limitations, in an amateur (i.e. not a professional) recovery situation, if nothing better is available to perform cloning, and if the risks of such DIY are accepted. Modifying the OS drivers to adapt their behaviour can also help, but that's a whole different skillset...

I hope someone with more time comes along and answers all your questions, but personally I think you might be expecting a bit too much with your level & quantity of questioning.

Thank you for your time and effort, Vulcan!

I am under the impression that most of the questions that I have asked can, by all fairness, be grouped together, as I am basically asking the same thing with 3-4 question marks - i.e., the question about PIO mode. So I think it might only seem that I am asking more, even though it's still that same question that I am not getting any form of an answer to. Taking that into account, 14 is something like 5-6. I guess the reason that I am writing extra is due to the somewhat evasive / abstract nature of some of the replies in this thread - almost no specific answer is given to rather particular questions (that damn PIO mode, lol). But as you state it - perhaps these points would take a rather demanding amount of time to cover, so I am happy with anything that I get as a reply, either way.

I have read the GNU ddrescue manual the very first time I attempted logical recovery and some after that, when I was drawing a comparison to dd over time. What I meant by asking "any other useful built-in algorithms" - is basically whether it had anything else up its sleeve against dd. My main intention here is clarifying whether gddrescue lives up to being perceived superior to dd in terms of the end result (amount of data recovered!). Here's a paste from the man:

GNU ddrescue manages efficiently the status of the rescue in progress
and tries to rescue the good parts first, scheduling reads inside bad
(or slow) areas for later. This maximizes the amount of data that can be
finally recovered from a failing drive.

The standard dd utility can be used to save data from a failing
drive, but it reads the data sequentially, which may wear out the drive
without rescuing anything if the errors are at the beginning of the
drive.

Other programs switch to small size reads when they find errors, but
they still read the data sequentially. This is a bad idea because it
means spending more time at error areas, damaging the surface, the heads
and the drive mechanics, instead of getting out of them as fast as
possible. This behavior reduces the chances of rescuing the remaining
good data.

Honestly, this seems confusing. I do not understand how is gddrescue supposed to recover any more data than dd? As it occurs to me, this is only valid if gddrescue gets really lucky on the X-th attempt (starting from a 2nd pass) of retrying to read a bad block. Does this operation make sense with a drive with a lot of bad sectors? I mean, this will cause a tremendous amount of time to complete due to the I/O error influenced timeouts (the benefits of using a large block size are negated by this). Now, how on earth does dd spend any more time than gddrescue at error areas? And how does the latter "get out of them as fast as possible"? I have witnessed the same amount of delay for either on 1st pass!

Overall, I am not asking for major details or anything that massively technical - just some pointers here and there so that I could do better research without wasting anyone else's time asking the same thing over and over.

Thanks!

[Vulcan]

Thank you for your time and effort, Vulcan!

I am under the impression that most of the questions that I have asked can, by all fairness, be grouped together, as I am basically asking the same thing with 3-4 question marks

I politely disagree - each question you ask is slightly different; answering one part doesn't (in most cases) completely answer the others, thereby leaving the thread open to questions about the parts not answered. In previous "theoretical question" threads, I have seen this happen, and it's been as total waste of my time. For that reason I'm not going to start trying to answer many of your (re-stated) questions here.

As I said before, IMHO most answers are "it depends on the specific behaviour of a specific faulty disk" - and I should add that also "it depends on your skill, in understanding and choosing from the various ddrescue options, as appropriate for that specific disk behaviour".

so I am happy with anything that I get as a reply, either way.

So here's just a brief reply.

My main intention here is clarifying whether gddrescue lives up to being perceived superior to dd in terms of the end result (amount of data recovered!).

If you don't believe that it is generally superior, then don't use ddrescue! Perhaps ddrescue might be too complicated for you to use it efficiently - its range of options can overwhelm some people.

I've given one simple example of a comparison below to try to help you see how ddrescue behaviour can help to recover more data than dd would on its own, but I'm deliberately not trying to explain all possible cases. You can find many other websites & blogs with ddrescue examples to read.

I do not understand how is gddrescue supposed to recover any more data than dd?

It seems you haven't considered some disk failure modes, along with other factors e.g. the risks of human error when manually using dd in multiple invocations, with different parameters in each invocation.

To give you just one short example - consider a 500GB disk with 10GB of sequential unreadable sectors at the start of the disk. On this disk, dd would be spending time sequentially trying those first 10GB (and therefore rescuing no data), whereas ddrescue would start to skip ahead after the first error and reach the later, readable, part of the disk much more quickly.

This ddrescue behaviour can be vital if the disk dies catastrophically after a short period of time - ddrescue could have recovered some or all of the readable data, whereas dd may still be attempting to read those first 10GB and so have rescued nothing before the disk died! I believe this is the scenario which Antonio Diaz was referring to, in the second of the three paragraphs you quoted from him.

Overall, I am not asking for major details or anything that massively technical

IMHO to properly answer some of your questions, yes it would require major details - you just don't realise it. Many answers would need to say something like: "in case X then [paragraph A], but in case Y then [paragraph B], and in case Z then [paragraph explaining something completely different]". Just because a question seems simple, does not mean that the answer is!

Good luck!

[HDD Spaz]

I can see this Q & A session resulting in a data recovery attempt. Will it be successful? Time will tell.

I plan to hand it over to a pro but I require knowledge on repairing the disk and the best way to clone it before hand it over.......

If DATA = Important, then be very careful.