Logical versus physical reason for bad sector development?

[SteinarN]

I recently discovered this forum as I was searching for information after a serious HDD crash on my HP nc6320 XP laptop with a ST98823AS 80GB drive.

The story is as follows:
The HDD was performing good as far as I could tell. Acording to SMART data retrieved by PerfectDisk the bad sector reallocation count was zero. The drive was slightly less than 50 percent full and system recovery was set to use 3 percent of the disk.

I then made the fatal decision to make a full backup of the disk using the factory installed HP backup and recovery utility. In one window the recovery utility said I could chose where to save the backup, i hit next and the utility started working. No possibility for choosing where to save the backup yet so I assumed the utility was only calculating the space requirement at this stage. This process took a long time. After 15 minutes it was still ocupied with what I assumed was space requirement calculation and I went out to do some work. When I returned an hour later I discovered to my horror the full backup now was saved on the C drive! And only 2 or 3 percent free space left! But i got worse. I could not find a way to delete the backup. I searched for an hour without finding a way to delete it. I then went on to disable my system recovery in a hope that this would allow me to delete the backup folder, but no joy. Next I reactivated my system recovery without knowing that it as standard set the allocated disk space for the recovery to 12 percent. As it was previously set to only 3 percent and the c drive had only 2 or 3 percent free space left it is obviously that there was not enough free space available to activate the system recovery at 12 percent disk allocation. I quickly discovered this and adjusted the space allocation to 3 percent, but only a few seconds later the desktop icons started to disappear, the computer froze completely within another few seconds and I had to force it off by holding down the on-off button. When I tried to turn it on it stopped very early in the boot process as it tried to load one of the registry files but couldn't find it. This is about a second or less after the bios leaves the control to the operating system in the boot sequence.

Upon further investigation I found the HDD to be impossible to read with explorer when connected as an external HDD to another machine albeit it was recognicable as a drive after some lenghty delay. I ran HDD Regenerator in scan mode to find 23+ bad sectors and about 2500 slow sectors. I then ran another recovery utility to save about 10 GB of important data. This took more than one day. Next i ran HDD Regenerator in repair mode. It allegedly now found and repaired about 2500 bad sectors. This took several hours. I don't know if it did something with the slow sectors though. Repeated this step once more and it allegedly found a few more bad sectors. The slow sector count was now much lower.

Now the drive was immediately recogniced by explorer, but the file system was in RAW format. I then formatted the drive and scanned it with checkdisk. No fault found. I went on to installing it in my laptop and have sucessfully installed XP again. The drive seems to perform flawlessly again except one minor detail. The bad sector reallocation count in SMART is now about 4500! This causes the bios to halt startup demanding to hit F1 and a message to immediately back up my drive. Very understandable demand considering the SMART data.

So, my question is what really lies behind the so called bad sectors. In this case it seems to be completely caused by some sort of logical errors and not any real physical condition with the drive or platters. And how could the sequence explained here cause the drive to be rendered useless?

Any comments?

Steinar

[Vulcan]

In this case it seems to be completely caused by some sort of logical errors and not any real physical condition with the drive or platters.

I politely disagree, based on many years' experience of looking at SMART data while working with drives. When you first ran HDD Regenerator (sometimes called HDD Destroyer ), you say it reported "23+" bad sectors - that was a clue that your problems were not only "logical".

However a detailed investigation would need information which you might not be able to supply. Do you have a copy of the full, raw, SMART data from before you ran HDD Regenerator? If that shows no problems, then do you have a time machine, to allow better diagnosis of that drive, before you ran HDD Regenerator?

Without those 2 things (or at least having the initial, full, raw SMART data and it showing confirmed problems), then any discussions here are just going to be opinions, with no ability to get confirmation. Other people here might have more time to involved, of course, so you might get more suggestions.

If you do have that initial, full, raw SMART data from the drive, from before running HDD Regenerator for the first time, please supply it.

[SteinarN]

I don't have complete SMART data. The only thing I remember is that I noticed the replaced bad sector count was zero a few days earlier when looking at that data in PerfectDisk. As to the other data like spin up time I don't know what those was.

The thing that puzzles me is that this fault happened at the exact moment I activated system recovery with an impossible setting for free space requirement. Something very serious happened at that moment. However, the cause for the error is only speculation at this time.

When I first ran HDD Regenerator the fault had happened and there was numerous bad sectors, of which many of them was at the very beginning of the disk which caused the laptop to be completely unbotable, so naturally, if HDD Regenerator can do any sort of bad sector scanning it had to find bad sectors at that time.

But, I have full SMART data now after the fault. The disk is fully operational. As I understand it, most data is incremental so it should give an upper/lower limit for what the data was before the fault happened.

Here is that RAW data. I have included only the three last digits where the preceeding digits are all zeros:
1: -9D6
3: -000
4: -1B13
5: -12F8
7: -164CBBB5
9: -9325
10: -000
12: -434
187: -FFFF
189: -26
190: -7819150019
192: -D8
193: -14FDEE
194: -1A
195: -4FC5A80
197: -2847
198: -2847
199: -000
200: -000
202: -010

[Vulcan]

I don't have complete SMART data. The only thing I remember is that I noticed the replaced bad sector count was zero a few days earlier when looking at that data in PerfectDisk. As to the other data like spin up time I don't know what those was.

And that is exactly the problem - those other values are needed (unless you have that time machine, that I mentioned ). This means that vital evidence of the original disk's state has been lost.

But, I have full SMART data now after the fault. The disk is fully operational.

See below for why I disagree with your assessment of the disk being "fully operational".

As I understand it, most data is incremental so it should give an upper/lower limit for what the data was before the fault happened.

I don't fully understand what you are saying. If you are saying that raw attribute values only increase, then that is not correct.

Here is that RAW data.

Perhaps I was unclear, sorry. I was looking for the full SMART data, but including the Raw attribute values in addition to the usual "value", "worst" & "threshold" normalised (or "cooked") values.

Although you only have the SMART data from the disk now (not from before running HDD Regenerator), and therefore it is of limited usefulness, look at the raw value of attribute #197 (Current Pending Sector Count) which is 0x2847 = 10311 decimal. I hope you can now see why I disagree with your assertion that the disk is "fully operational"!

[SteinarN]

When I said the disk was fully operational I meant that it don't show any sign of failing except if you look at the SMART data. However I don't trust the disk the least bit and I have already ordered a new disk.

I did not say the SMART data only increases, if you take the time to read what I actually said you will see that I said most values is incremental. Of those 20 values I mentioned are about 12 or 13 incremental and only inceases in value as far as I can estimate.

What I wanted input on was the reason for the crash and any possibility for a connection with what I did, why it crashed so spectacularly and quickly just as I activated the system recovery, putting the system in an impossible situation regarding the free space requirement.

[Vulcan]

When I said the disk was fully operational I meant that it don't show any sign of failing except if you look at the SMART data.

I expect that there are other signs, not limited to the SMART data, but this investigation would require you to do more diagnostic work.

However I don't trust the disk the least bit and I have already ordered a new disk.

Good plan

I did not say the SMART data only increases, if you take the time to read what I actually said you will see that I said most values is incremental. Of those 20 values I mentioned are about 12 or 13 incremental and only inceases in value as far as I can estimate.

I did read what you said - it was unclear to me about exactly which SMART data you're referring to, what "upper/lower limit" you proposed could be inferred and from what data etc., which is why I tried to guess what you meant. Anyway, since you're criticising my ability to interpret your comments, obviously I'm not the right person to help you, so I'll leave you to figure out the rest. Glad you're getting a new disk.

[SteinarN]

I don't have the ability to figure out much more than I already have. The discussion I hoped for was if bad sectors is ONLY a result of the drive quality and health or if bad sectors can be software induced in rare cases.
The other thing that puzles me is why so many bad sectors could appear apparently simultaneously and no more bad sectors in the probably ten hours the drive have been spinning since I got it going again. But then again, I'm just an ordinary guy which happens to be extraordinarily interested in a lot of different technical stuff.

Edit:
For the records, current bad sectors, 0x12F8, 4856.

Last edited by SteinarN on May 1st, 2012, 13:58, edited 2 times in total.

[BlackST]

The disk is NOT fully operational, is in "time bomb mode" ....
The use of "that" program was a bad idea.
There's no alternative to PROFESSIONAL refurbishing but it requires know how and tools.

[Vulcan]

The discussion I hoped for was if bad sectors is ONLY a result of the drive quality and health

No, although that is the most common cause.

or if bad sectors can be software induced in rare cases.

In your case, no. Readers with deep knowledge will understand why I said "in your case", but it's not relevant to your situation.

The other thing that puzles me is why so many bad sectors could appear apparently simultaneously and no more bad sectors in the probably ten hours the drive have been spinning since I got it going again.

There is more to this topic than where you are looking, which is why it seems puzzling to you. Perhaps someone else will have the time to explain more to you about how this functionality works.

@BlackST - All agreed!