3

What I mentioned as "bad sectors" here are quoted from following scan/test:

  • result from CHKDSK, referred as "Bad Sectors" on the report screen.
  • result from HD Tune - Error Scan, referred as "Damaged Blocks" or red square.

I have an internal HDD that has been successfully recovered from bad sectors by performing a full erase on whole sectors at that disk. (After an erase, running surface test told that all sector are OK, I don't see bad sectors anymore).

After that I encountered BSOD. After restart, I do the surface scan again, but this time 1 bad sector is found. I tried to erase it again, then running surface test told that the bad sector went away (again).

I already heard several times that sudden power-off or any power-related problem can break your hard drive. But can BSOD cause bad sectors, too? Or does this mean that my drive is already falling and need to be replaced? FYI, the S.M.A.R.T status from HD Tune is still "OK".

Since BSOD is quite common to occur (on my case, for every 3-6 months), I found it's annoying if after that I should repeat the "rescan the disk, erase, and restore" cycle.

EDIT: In reply to all comments & answers here

I found many article explaining that bad sectors aren't always physical reason. There is also logical reason (e.g. the CRC for that block does not match the data read back by the disk).

A logical — or soft — bad sector is a cluster of storage on the hard drive that appears to not be working properly. The operating system may have tried to read data on the hard drive from this sector and found that the error-correcting code (ECC) didn’t match the contents of the sector, which suggests that something is wrong. These may be marked as bad sectors, but can be repaired by overwriting the drive with zeros

Since there are too many, instead of copy the content, I will just put the link here:

2 Answers2

8

A Windows BSOD cannot cause a bad sector, if by that you mean a sector that will later read back as "Uncorrectable ECC error" or similar. This is because software has no access to anything at any lower level than the drive's connector, and there is ordinarily nothing that can be done at the interface of a properly-working disk drive that can cause a bad sector. Not under Windows.

I say "ordinarily", "not under Windows" because the Unix-y utility hdparm does have a --make-bad-sector option. It works by invoking either the WRITE_LONG or WRITE_UNCORRECTABLE_EXT ATA commands. But the Windows disk drivers (we are talking about Windows, since you're talking about a BSOD) provide no way to request these functions.

You likely have a failing hard drive. You had one bad sector (that you knew of), but since then it developed another. That isn't uncommon. The subsequent errors may or may not have caused the BSOD. I can imagine several HD errors that could lead to a BSOD, most likely due to failure to resolve a page fault that was raised in kernel mode (BSOD code KERNEL_INPAGE_ERROR).

Incidentally, a BSOD cannot be caused by a third-party app. Only by buggy kernel mode code (like a driver), a bug in a protected system process, or failing hardware. Bugs in apps can only cause app failures. And BSODs are not at all "common to occur" on properly working hardware with reliable drivers.

2

The typical traditional terminology for a "bad sector" is a portion of a hard drive that is physically failing to work right. It's just plain broken. That is why you're getting some people reporting that a BSOD shouldn't cause this. A BSOD may cause invalid data to be written; some advanced RAID cards have been known to be battery powered so that they can be more sure to write out a complete block if they lose power. (These batteries may be a rather pricey add-on to the base cost of a RAID card.)

When Chkdsk is reporting a bad sector, it really means that Chkdsk had a problem; this can be caused by logical faults, which may be different than a hardware-based "bad sector". The reason it sounds like you're reading two different things is that the term "bad sector" has been used to describe two different types of problems. Advice discussing the concept of a physically bad disk might not apply very well to the topic/concept of a disk that has an invalid filesystem volume caused by corrupt data being written, which might happen if the operating system stops running normally (like when a BSOD happens).

Most BSODs that I've encountered did not result in data errors; I tend to get concerned about hardware reliability anytime I find Chkdsk is reporting bad sectors. Your best bet is to not rely on the drive as a sole way of storing any important data. That's always a good idea (important data should be backed up). If you're one of the more-than-2% (*) of people who haven't bothered to implement a proper backup solution, then this may be an excellent reason to stop allowing excuses to convince you to do things more risky.

(*) I know, it's much more than 2%. Probably way more than 3%, as well.

TOOGAM
  • 16,486