My shitty week - Bad Sectors on Hard Drives.

[Novaer]

So I've had a rather interesting week which has caused horrible slowdown on my end and delays in fixing a few things. My computer had not one, but two hard drive failures, and both from my WD Black Drives! ?If you haven't experienced such a failure; then it's a very tense few hours - or days depending on how big/complex the arrays are; mine wasn't that complex but on the of the drives that failed was my newly acquired WD Black 4TB Hard Drive. 

If you aren't aware of how nerve wrecking it is; check out this video by Linus Tech Tips (A fantastic and well informed channel by the way) which shows the failure of their Whonnock Server and gives an insight into the stress of these events. At first mine was no where near as drama filled as the poor guys over at Linus Tech Tips, I don't have my drives in any kind of array, or anything like that. They're just as they are in Windows, their own drives no RAID or anything that like that. 

So getting to it; on Thursday Morning my first indication that something was up, was me booting up Windows and seeing this:

"Okay" I thought, "No big deal. Probably just some corruption or something. Sometimes shit like this happens. My E: Drive is my Steam Drive so if I have lost anything then I'll just have to redownload all my games." So I patiently waited three quarters of an hour for it to complete before I could even get into my computer and start working on EcchiDreams... Okay that's a lie; I just went away from the computer grabbed a coffee and worked on EcchiDreams from my phone. xD

After it was done and I logged in, I accessed it and everything seemed fine. Until I got alerts telling me that I needed to restart my computer and let; what I assume is a really dumbed down chkdsk (Check Disk) to run again. 

Erroneously thinking that perhaps Windows knows what's best for my hard drive, I allow it to reboot and let it go through, once again, nearly three quarters of an hour to do what it was supposed to have done the first time around. Afterwards I logged back in, and again - it asked me to restart to submit to another round of CHKDSK'ing. "Fuck. That. Noise." I thought. 

This time, I decided to run 'CMD' in Administrative Mode, so I could run 'chkdsk e: /F /R /X /B' in the standard mode I am used to.  To break this down into people who might not be aware of what this does:

• chkdsk - The name of the program used; stands for Check Disk. It's a Microsoft program that's been around for donkey's years. Not to be confused with MS-DOS Chkdsk, which worked, I think, along side ScanDisk back in the days of computer's medieval years.
• e: - The 4TB Steam Drive.
• /F - Fixes errors on the disk (Otherwise CHKDSK will run in read only mode)
• /R - Locates bad sectors and attempts to recover readable information. It implies /F if you don't use /scan (which is an online scan)
• /X - Forces the drive to dismount (Become inaccessible to Windows Explorer whilst CHKDSK is looking at it) I had to close Steam too.
• /B - Re-evaluates bad clusters on the drive, which implies R. 

I used F and R because of habit. I want no room for misinterpretation. I might not be worried about the contents of the drive but this is a brand new drive. I only got it a few months ago, and it's hardly ever been used.

After sixteen fucking hours it almost completed. It just finished Stage 5 and then it crapped out. 

Quote

C:\WINDOWS\system32>chkdsk e: /F /R /X /B
The type of the file system is NTFS.
Volume dismounted.  All opened handles to this volume are now invalid.
Volume label is Steam Games.

Stage 1: Examining basic file system structure ...
 359168 file records processed.
File verification completed.
 0 large file records processed.
 0 bad file records processed.

Stage 2: Examining file name linkage ...
 451310 index entries processed.
Index verification completed.
 0 unindexed files scanned.
 0 unindexed files recovered to lost and found.

Stage 3: Examining security descriptors ...
Security descriptor verification completed.
 46072 data files processed.
CHKDSK is verifying Usn Journal...
 325152 USN bytes processed.
Usn Journal verification completed.
Removing 2 clusters from the Bad Clusters File.

Stage 4: Looking for bad clusters in user file data ...
 359152 files processed.
File data verification completed.

Stage 5: Looking for bad, free clusters ...
 676614717 free clusters processed.
Free space verification is complete.

A disk read error occurred c0000185
An unspecified error occurred (6c6f6766696c652e 29f).

At this point I'm thinking "Oh fucking hell, no..." and go into full investigative mode. So the first thing I wanted to do was look into the S.M.A.R.T (Self-Monitoring, Analysis and Reporting Technology) Data. S.M.A.R.T is usually built into most if not all hard drives nowadays, I also decided to download a set of tools to get as broader of a picture as possible, and to work within those programs hence why the screenshots will use different programs. 

If you've stumbled on this topic because you're experiencing the same problems, the programs I downloaded were:

• Crystal Disk Info 7.0.3:- http://crystalmark.info/?lang=en - Shizuku Edition, Ultimate (Portable)
• Western Digitals, Data Lifegaurd Diagnostics 1.29.0:- http://support.wdc.com/downloads.aspx?lang=en - Remove all flashdrives before running this, as it crashes if you do not.
• HDDScan 3.3.0:- http://hddscan.com/ - Pretty damn good software in my opinion.
• Arconis Drive Monitor 1.0.0 (Build 556):- http://www.acronis.com/en-gb/homecomputing/download/drive-monitor/ - Monitors your drives SMART read outs and tells you in advance if any of your hard drives is about to fail. 

First things first I ran Crystal Disk Info, and found out that it wasn't just my E: Drive playing up, but something much darker and more sinister was started to gloom over the horizon. My F: Drive as seen in the screenshot below which believe me - we will get to later:

So as we can see here there were 3 sectors reporting under C5 - Current Pending Sector Count. You might be asking, "But Wolfie, what is a C5 - Current Pending Sector Count?" Wikipedia has a record that is better than my shitty words can describe it, but an article I have to hand can fill in some blanks afterwards:

Quote

Count of "unstable" sectors (waiting to be remapped, because of unrecoverable read errors). If an unstable sector is subsequently read successfully, the sector is remapped and this value is decreased. Read errors on a sector will not remap the sector immediately (since the correct value cannot be read and so the value to remap is not known, and also it might become readable later); instead, the drive firmware remembers that the sector needs to be remapped, and will remap it the next time it's written.

However, some drives will not immediately remap such sectors when written; instead the drive will first attempt to write to the problem sector and if the write operation is successful then the sector will be marked good (in this case, the "Reallocation Event Count" (0xC4) will not be increased). This is a serious shortcoming, for if such a drive contains marginal sectors that consistently fail only after some time has passed following a successful write operation, then the drive will never remap these problem sectors.

So a pending sector isn't one that is necessarily bad. It could (remote possibility) be a case that if we had a power cut whilst it was in the process of (less likely) reading, or (more likely) writing to the disk. This could be exaggerated if I have Write Caching enabled (And I do). These come in two shapes and sizes, hard bad sectors and soft bad sectors, this article explains more on it (Archive.is - How-to-Geek).

Quote

Slightly Off-Topic | You cannot Low-Format modern hard disks anymore, back in the day there were three types of formats:

• Low - Retracks the Drive, useful for really old drives when format ceases to work properly. This cannot and should not be done on modern hard disks (Seriously, at best you'll be wasting your time. At worst you'll really screw the pooch.)
• Quick - Deletes the Partition and replaces it with a new one. Data on the disk isn't actually destroyed, which is why one must never just quick format a hard-drive before selling it, under any circumstances. It can be recovered relatively easily. Back in the day this was a bad practice with much older OS's.
• Full - In a world of 1's and 0's it simply overwrites the disk with all 0's. Essentially, it gives you a completely blank drive. Although in theory, with older drives, some people might still be able to recover data from this if they really wanted to, although it's a lot harder or even impossible in newer drives according to some sources.

Which is where Western Digital's program comes in handy, as I explained earlier: Drive (E:) is only my Steam Drive; I can easily re-download everything in Steam again if I format it, so I chose to have the Western Digital Software, to format the Western Digital Drive, remove all flash drives before using it, because otherwise it will crash before it even loads.

I could have gone for Microsoft's formatting tool but given how it failed at the last moment with CHKDSK, I wasn't going to take that chance.  At first it said that everything was fine, the SMART Data was giving me different readings to the SMART Data in Crystal Disk- which was confusing:

"O...kay..." I thought to myself. Not satisfied with this I decided to launch a 'Quick Test' only that failed rather spectacularly. 

Quote

06-Quick Test on drive 1 did not complete! Status code = 07 (Failed read test element), Failure Checkpoint = 97 (Unknown Test) SMART self-test did not complete on drive 1!

"Not so perfect anymore, eh, Western Digital?" Followed by several swear words, because I like swearing. Calmly I opt to do what's called a 'Write Zero's' run which is basically a full format: 

This was nearly three hours in, that I took the screenshot, with just over four hours remaining. Yes it took six to seven hours to complete. By now it was ten in the morning on Friday, I had been working on this all night long. So when I got up on Friday evening, I ran Crystal Disk Info's Smart Tool again and got this:

So not only did I get rid of all C5 - Current Pending Sector Count Errors (WOO~!) I also avoided C4 - Reallocation Event Count and 05 - Reallocated Sectors Count, which meant they were indeed soft errors which a format fixed perfectly.  So ... YAY! ?

But remember that dark and sinister cloud glooming over the horizon, that I mentioned earlier? Well, in my effort to work on one problem at a time I overlooked something very, very critical. My (F:) Drive. So I took a look, and what I found floored me. My (F:) Drive is a very, very reliable drive, unfortunately it's also very, very well used and old. Let's put it this way: I got it in late 2009, brand new. Originally it stored my Operating System and it was the drive I used on a daily basis; eventually it got designated to my NAS drive for a year or two, before making a come back in my PC around about the time I started number crunching on behalf of Folding@Home. It had been checked for errors and was doing relatively fine considering it's age, it certainly appeared healthy that's for sure. 

At the end of 2014; I started using it again to store the EcchiDreams backups; as well as all my working data - any files I modify for EcchiDreams and so on, as it had plenty of life left in it, it also held my WAMP set up for developmental testing as well as various bits and pieces critical to EcchiDreams. Not only that but it it housed a second partition which had all my backups on. It wasn't going to be excessively written to, but it wasn't going to be doing sweet fuck all, either.

Now these programs were telling me that it was in trouble... How much trouble I didn't know; until I clicked on it. 

BIG PROBLEM and at this point I'm at red alert. The 05 was telling me that at some point it had already reallocated a single sector, however C5 was telling me that 493 were pending to be reallocated. This wasn't even the half of it. 

I was thinking "My F: Drive" also a WD Black, an old 640GB drive "Was on it's way to live up to it's drive letter - Fucked... Maybe I'll get lucky again. Maybe these are soft bad sectors. Maybe I am over reacting, after all! The 4TB had some C5's so maybe they're just power failure related counts too?!" 

Oh how wrong I was... At this point I ripped everything off of it and essentially backed it up which took some hours and a number of files didn't make it, as Windows was complaining that it couldn't read from the source. These files thankfully were redundant, temporary files or backups - of which I just replaced when I moved the data off of it. 

I started up HDDScan, and took a look at what was going on in the drive. For around 46% of it Then I started getting this:

"Well fuck..." I thought. The drive had been scanning between 100,000 - 180,000 KB/s but in this area the speed dropped to 256 KB/s. After a few hours this completed and showed up hundreds of bad blocks. So I wiped them over with all Zero's and thought I'd take a look at it after doing another read test. Getting some kind of background logging was important to me; so I went ahead and downloaded/installed Acronis Drive Monitor (Which is free) and let that run in the background.

After this I went back to do another read test to find that it found more bad blocks, "Oh shit..." I thought, and so I wiped it again and repeated the process... Which is where we come to now, fucking Saturday morning!

Needless to say; it was clear now that the problem was getting progressively worse. This is NOT what you want to see in a Hard Drive:

934 Reallocated sectors, and that number keeps going up every time it's written and read from. I'm no expert in this, but I assume when a hard drive does this it means it's failing, and absolutely should not be used for anything considered even remotely important, if at all. If your hard drive is failing in this manner you will want to get as much of your data off of it as soon as possible, ideally you should have backups. I suspected this was happening and so I took as much as I could (which thankfully was almost all of it) from the offset.

I am extremely lucky that my 4TB HDD fucking around caused me to see that there was a major problem brewing on my 640GB HDD; if the 4TB one hadn't have been screwing around when it did, then I probably wouldn't have noticed the problems with the 640GB one until it was too late. Like a Canary in the Coal Mine. Thanks to the 4TB HDD's shenanigans I actually saved my data. In the next post I will be doing a post-mortem, if you will of the 640GB HDD. That means taking the bastard apart and looking inside. I doubt we'll see what caused the fault as it could be anything (I will explain in the next post) but this is EcchiDreams, and stripping down a hard drive sounds like it'd make fantastic computer porn.  I will reserve the post until then. :)

[Novaer]

Inside the Failing Hard Drive

Notice: I've compressed the images as best I could as we don't want slow loading times, unfortunately this could have compromised quality. [Used: https://shortpixel.com/]

This breakdown will also have a running commentary that I hope to provide some educational material, something to read as I go through each picture, and hopefully explain all the things you see in the picture above. This is of course if you want to read it or are curious about anything. Of course; you can always leave me a comment, and ask, someone might answer for me if they know, but if not I will happily answer any questions. I am also not a hardline expert in this - so if I get it wrong do let me know and I will make the appropriate corrections with attributions. 

Before I opened the drive, I took off the PCB on the back (The green board).

This board is a 'controller' of the hard drive, it's where the processing takes place, the board is [2060-701590-000 REV: A, Type 1 - Without BIOS]. There are three notable chips on it:

• CPU (88I8846E-TFJ2) - [Couldn't find Datasheet]
  • I think it's a 64 pin, (64 pins input and 64 output) IC Controller for the Hard Drive. I think; it's made by Maxwell/Marvel and according to some sources it's a dual core 600Mhz chip. Rather respectable for a harddrive of its time.
• Cache (K4H561638J-LCC, Samsung 910 Chip) - [Datasheet]
  • A 16M x 16 chip (256 Megabits = 32MB) Cache for the Hard Drive. It's a DDR400 with a CAS/CL of 3.
• Motor Spindle Controller + VCM controller, ST Microelectronics (L7251 3.1, SMOOTH) - [Couldn't find Datasheet]
  • I assume this ship controls the spinning of the platters. 

There is space for a BIOS chip; although it doesn't seem to come with on installed, my theory is that the firmware for the device is build into the CPU; or something. This board seems fine; after all if it wasn't I wouldn't have had the failures that I had, it would have just outright stopped working.  Another thing I noticed is that the once golden contacts were tarnished and some were even outright discoloured.

From assumption here; but once the data has come through the SATA Connectors on the front of the board, and is processed by the onboard chips and such, it's then sent out through these pins to the heads, which we'll see further down the page. 

Now I'll open the hard drive... 

STOP! If you're thinking about doing this to your own hard drive; and it's in perfectly working order and you value the data on it, then do not open it up. This will pretty much fuck your hard drive over. Sure; it might work for a little while but it won't for very long after opening it up. This hard drive in question will be destroyed afterwards through other destructive means as well, so I do not care about it which is why I am opening it up.

So inside the hard drive is what you'd typically expect. Up there in all it's blurred glory is a filter, this is designed to catch particles of dust/metal or what have you that might come off the drive as it spins in a clockwise fashion, at 7200RPM anything sent off by the drive could potentially damage it, under normal operations, at least that's what I'd assume anyway.

There are two thick platters (I've not seen platters this thick for a very long time) and they seem to be solid metal. I have seen glass platters before (In a laptop hard drive that had failed. Upon opening it up it was clear that it had been shattered in a billions of pieces) At first glance as I predicted there was nothing (visible) out of the ordinary on the platter.  There are rumours on the internet in the HDD circles that this model of Hard Drive has one platter - This as evidenced by this picture; is categorically false. I suspect each platter is about 320GB, about 160GB per side.

Moving down the image, we see the Head Armature, which houses what I assume to be four heads. These are extremely tiny and extremely sensitive little things.  These work by hovering over the drive whilst it's spinning, it's rather complicated but it's one of the reasons you should never open a drive, especially in your living room like I did, if you want to use it again. It is a common misconception that inside of the hard drive is a vacuum, this is false. It's actually highly pressurised and it's this pressure coupled with the disk's high speed spinning that causes the head to glide over the platter. When you fuck with the contents of the drive (IE the, Air Pressure) it could be enough to cause what's called a Head Crash, or worse. They are incredibly sensitive pieces of equipment how they work largely depends on the technology employed this is to get the density that is often required with modern hard drives. I cannot remember off the top of my head what the different types actually are. These heads are capable of reading and writing at the same time and use exceptionally teeny-tiny wires that travel from the SATA port, across the chips mentioned earlier, through the dirty contacts I mentioned earlier through a ribbon cable up to the heads. The heads move across the platter thanks to a magnetic actuator and what I assume is current that allows the head to move with extremely precise and controlled measurement.  All four heads seemed to be in perfect working order, again the type of errors I was seeing was indicative to a head failure, at least not an outright head failure. They were all mounted properly into the head parking zone (Pictured Below) as they should be. The heads in theory should remain here until the drive is spinning at the correct speed for them to just glide over the platters. 

On the other side of the head armature we have a really cool piece of technology that makes the armature move across the platters. It is what I briefly talked about - the actuator, and the actuator coils. Those massive metal blocks you can see are extremely powerful magnets, of which I do have a very funny story to share with you about from 2006 (Nearing my end of Secondary School). My ICT Teacher (Who had no idea what they even were) took one of these off of me, because like a sixteen year old should with an idiot ICT Teacher who couldn't tell a Hard Drive from a Motherboard from a Desktop Computer (Yes, seriously) I was being a little unruly shit and fucking up the monitors before degaussing them with the inbuilt features (I was bored) and trying to give my friends a better education than the teacher was giving. He came up to me, and confiscated it, put one of them in his pocket and sent me out of the class. What happened was something I had to be told about by him; because I wasn't there.

This guy loved to lean against things, one such thing he leant against was a (and I shit you not) a metal filing cabinet. After class he asked me just "What is this?!" to which I replied "It's a Magnet." He replied "Well no duh... Where did you get it from? Magnets aren't usually this powerful and I'm concerned you've stolen it from something industrial, at school." He asked and then asserted to which I replied with a defiant, questioning look "If you knew anything about computers. You'd know where that magnet is from, and I resent your accusation. It's from a hard drive, MY hard drive, a failed Seagate one-" And I explained what it did and why every hard drive had one, because the cheeky bastard had the balls to insinuate I was being untruthful. So I educated my educator (Which I think a five year old could have done).  Anyway - He told me that after I had been sent out he leaned against the filing cabinet and had become stuck - because he didn't want to ruin his trousers, and told me that I almost had to pay for a new set of trousers, to which I told him that he shouldn't have been so stupid as to stick it in his fucking pocket. Apparently it had caused a lot of amusement for my classmates who told me that he was panicking when he had gotten stuck to the cabinet. I wish I fucking saw it. 

Anyway... Back on topic. These things are fucking powerful, I think they're Neodymium, but I am not 100% sure. The coils:

I assume get changed with a certain current in a certain way, and this causes the coil to move along the magnet which moves the arm. A modern hard drive can do this with scary precision, in (usually) under 50ms, as specified by the manufacturer these drives can usually locate the section of the drive you're looking for in under 10ms (8.9). So it's very fast and extremely precise. Normally having magnets this close to the hard drive platter is a recipe for disaster. Whilst the platter itself isn't magnetic, the coating on it is - not enough for these things to stick to it mind, but definitely enough to wipe out all the data on it, but these magnets won't do that because they're contained and stick - floating by the mounts, to each other keeping the field contained, and I assume the stronger the magnet, the better the resolution for the movement of the head. If you do take a hard drive a part - be careful when you remove the magnets, but make sure you salvage them because they are hella fun. I say be careful because you can and will hurt yourself if you're not. 

The ribbon cable you can see going into the PCB on the side of the arm comes from the dirty contacts I mentioned earlier.

The PCB is about the side of my thumb from tip to first joint. Under the ribbon this is what it looks like:

Oh yeah, and I fucked up the heads (to get some close ups).

Before I removed the assembly from the drive, this is what the heads looks like when parked:

Before I took the heads apart, I used my mobile phone's light to shine down on the drive, this is where I discovered something along the surface of the drive. I have no idea if I had caused this, or if this was the errors my drive was facing. This is the last picture I have on this, and as you can see there is clear damage on the edge of the platter (That I moved around so it could be near the heads). I did not move the heads onto the platter so I have no logical explanation as to where these scratches had come from. 

Now here, I am not exactly sure as to what had caused this, and it took shining a light from a specific angle just to see it whilst I was slowly spinning the to look for damages and take pictures, I didn't do anything other than open the drive up at this point. Honestly I wasn't expecting to see anything, at all as the damage could have easily have been invisible, magnetic damage. Such as weakness in the actual coating, that constitutes the sector. I am 99% percent sure that this was there before I opened the disk because I did not touch the platters at all. However there are a few things that I think rules that the bad sectors are because of this; but again I am not expert:

• I've known hard drives to write from the edge and work there way in, and I know CD's that write inside and work their way out, so I am uncertain if this is the case.  Assuming that each side is 160GB (Each platter is 320GB), and assuming that the top side is 'Platter 1, Side 1' this would be 0% - 25% of the drive. In this case I'd expect it on the second side of the first platter.
  • Even if it was 'Platter 2, Side 2' then it'd be on the 75%-100% range of the HDDScan.
  • The damage was in the 42-48% range.
• The bad blocks were appearing after each read following a write of all zeros. This suggests magnetic failure (The coating is no longer retaining information or working correctly). This is clear, physical damage. 
• There would have been shrapnel, and the only thing that I can think of that would cause this is a head crash, or even the arm clipping the edge of the disk. 
• The platter couldn't have been spinning that fast for this if at all. If it was, I'd expect rings around the edge of the disk. 

Possible causes:

• The 1% of me that isn't sure - A Wolfie Error (An error made by me that I am not aware of) - this is most likely. But as to what - I am not certain. 
• Manufacturing defect that was programmed out of the drive - All (or almost all) drives have defects in them, it's the nature of getting this precise, nothing in the universe is perfect. The manufacturers usually have these zones mapped out of the drive so they're never used, although I think if its over a certain percentage the drive never makes it to market. All drives also come with reserve sectors so the dead sectors are merely remapped to these (like I mentioned in the previous post). 
• Another error that I didn't know about or was outside of the track range. It stands to reason that the head doesn't go to the absolute edge. It's possible that this is a defect but it was never detected because why would it be, if the heads simply do not go out that far during read/write operations. 

So I hope this was interesting... Or whatever. ^_^ I would have loved to have taken the platters off, unfortunately the last screw was in so tight that my screw driver (despite being the correct size) chewed through the teeth of the screw head, making its removal completely impossible.

[Neptune]

Very educational :D And very detailed. 

I love seeing computer parts like this. I always forget just how complicated and delicate these things are, and it always fascinates me to look at them in this level of detail. Thank you for sharing this. I know I've been sitting here for the last couple of days while you've been dealing with this, and sometimes helping you with the pictures, but I like the way you did the post. Informative and yet even someone that isn't that computer literate, like me, can understand. (Although I do know a little more than the average Joe, thanks to what I pick up from you)

[Novaer]

Thanks Neppy. Computer hardware is something I used to be very passionate about. I also love taking stuff apart. ?

[Whoreo]

I have to admit I didn't expect to see this kind of content here on EcchiDreams but as you said "This is EcchiDreams, and stripping down a hard drive sounds like it'd make fantastic computer porn. " XDD I found it to be very interesting as I have never seen inside of a harddrive before, I've never even thought to even Google it. thanks for this Wolfie I found it v. informative. although if you ain't no expert at this stuff I'm a complete dumb ass when it comes to this stuff. if I saw my hdd failing like the 4tb one you mentioned earlier I would have just scrapped it and got a new one.......

so the 4tb one was the one windows was complaining about; however that was fixed, but the one windows wasn't complaining about the 640gb one was the one in worse condition? that sounds like typical Microshite to me.

[Novaer]

Thank you for your glowing feedback. :D Essentially the TL;DR version of the story is indeed this:

•  My 4TB Harddrive had issues; that Windows detected. That's what alerted me as to what was going on.
• I investigated and fixed the minor fault with the hard drive. 
• In the process of this I discovered a much more serious fault brewing on my 640GB Harddrive that Windows was silent about.
• Turned out the 640GB was the one actually failing. If Windows didn't pick up on the 4TB drive problems I wouldn't have known about the 640GB drive until it was too late.
• 4TB was a canary in the coal mine.

That's why I posted this up, because I thought the story was fascinating, if a little strange and unusual too. I lost a hard drive, but now I can look back on it and it wasn't such a shitty week after all. I'm fucking lucky thats for sure.

[Neptune]

After watching you doing computer things for all of these years, Wolfie, I have learnt something about computers. They can fail in some of the most fascinating, unusual, frustrating and unpredictable ways possible.

One would think that due to the precision in which they are made, they would fail in predictable ways. Hell nawwhh! Tch. Technology xD The more simple it makes life, the more complicated it makes things lol

[Novaer]

13 minutes ago, Neptune said:

After watching you doing computer things for all of these years, Wolfie, I have learnt something about computers. They can fail in some of the most fascinating, unusual, frustrating and unpredictable ways possible.

Yes; exactly.

13 minutes ago, Neptune said:

One would think that due to the precision in which they are made, they would fail in predictable ways. Hell nawwhh! Tch. Technology xD The more simple it makes life, the more complicated it makes things lol

The problem is the more precise you get the more sensitive they get towards faults and damages. It's like people usually say 'They don't build them like they used to' I can agree to that to a degree:

I had a 12GB Fujitsu HDD that lasted near on twenty years before I got rid of it because there were better capacities. This 12GB HDD was still working when I decommissioned it, which was still in perfect condition. 

But the problem is if they built the 640GB one 'like they used to' it almost certainly wouldn't have been 640GB. This is the problem (kind of) today. Parts can be simply extremely sensitive to damage, which is why I wasn't expecting anything visible on the platter - because chances are; even if there was physical damage, I would have expected it to be microscopic or not even visible, hence why I was confused to see that kind of damage on the edge of the platter. Chances are the damage was caused by the magnetic media becoming less able to store changes in the bits that it stores, this damage is invisible to humans, usually. 

Although I really don't know what those scratches are, and I am skeptical that they are the cause of the failure; simply put - I've never seen anything like it at all, in my life.

[Novaer]

Update:

My first thought is that maybe this is some kind of SATA controller failure.  So I'm going to check this out. But here we go again... 4TB (Steam and Data Drive) is down. 

[Whoreo]

How did it go?

[Novaer]

I zero'ed the drive again as I don't really have the luxury to sort anything else out at the moment, too busy for that. But I still want to test the controller and the SATA cables, it's just finding the right time to do it.

[Novaer]

Update: I forgot about this topic.

Yeah that 4TB drive was trashed in the end, same issues. I'd format it (Zero the drive) and the issues would disappear completely, then I'd use the drive for a few weeks, before low and behold the errors in the same place started appearing again, but only when data was written to that area, and needed to be read back from it. These errors started going up and up, each time I zero'ed the drive, it'd all be fine and the errors disappeared, but when they came back, there was more of them, every single time.

I think in the end it got to the stage where there was around 10,000 of them or so and I decided - yeah, no. That's coming out. I got a new 4TB drive that's worked fine ever since (Same brand, model) and the old one was scrapped. Like literally - we turned it into a small ingot of Aluminium, or some bullshit like that.