Bad Sectors

[seekermeister]

I had always thought of bad sectors as being physical damage, and therefore cause for an RMA, but I found some old forum threads on the subject, one of which said this:

bad sectors are usually caused by cross-linked files which are a result of the operating system overwriting a portion of one file with another file. Potential causes of cross-linked files are turning off or restarting your computer without exiting Windows or by not regularly defragmenting your hard disk. Another reason is when the heads make contact with the hard disk platter(s).

My computer has had more than it's share of forced restarts, due to freezes where the only alternative was to use the reset switch. I knew that this could cause corrupted data, but never thought that it would also cause bad sectors.

If this is true, how would one determine if the problem is of a physical or a software based origin?

 

[speedgamer01]

Run a 'Defrag & Optimize' programe with the 'Auslogics Disk Defrag' free defrag software. (this run is a humble defrag run, it is just more complicated) After that your bad sectors should have gone (or moved) if they were software based errors.

 

[Ztruker]

Where did you find that quote?

I don't believe it's true. A bad sector has always been defined as a physical defect in the surface of the drive platters. The only fix I've ever seen that worked was to mark it bad so it's never used again. That's what Spinrite did.

 

[seekermeister]

Most of what I found Googling were forum threads, therefore not necessarily definitive:

bad sector causes - Google Search

This statement from Wikipedia:

A bad sector is a sector on a computer's disk drive or flash memory that cannot be used due to permanent damage (or an OS inability to successfully access it),

seems to allude to it, although it is not clearly defined.

Bad sector - Wikipedia, the free encyclopedia

The first paragraph of this article seems to confirm it:

http://www.datarecovery.com.sg/data_recovery/bad_sector_on_hard_disk.htm

 

[Teerex]

A bad sector is a sector on the drive that couldn't be read when it was called into service. The drive will attempt to read it again (or the defragmenter). If it is unreadable once again, it will be marked bad 4ever.

The drive the has two available options, depending on its overall condition and age - replace the bad sector with the backup sectors every drive always has - and if it can do that, the bad sector doesn't appear in the count (but you can see this kind of event in SMART under a different category).

If it has run out of backup sectors, you get a bad sector you get to know about.

 

[seekermeister]

As far as the "spare sectors" that are provided for this purpose...where are they? Are they interspersed everywhere, among the regular sectors, or in a block set apart from the others? If the latter, that would be the cause of possible fragmentation. Once that a sector is marked bad, is it totally ignored in the future by diagnostics, or will it always be shown as a bad sector? If the latter, then there is no direct means to know if it has actually been remapped or not. I suppose that the only way to do so, would be to totally reformat the drive, and see if the bad sector continues to appear.

 

[Bill2]

As far as the "spare sectors" that are provided for this purpose...where are they? Are they interspersed everywhere, among the regular sectors, or in a block set apart from the others? If the latter, that would be the cause of possible fragmentation. Once that a sector is marked bad, is it totally ignored in the future by diagnostics, or will it always be shown as a bad sector? If the latter, then there is no direct means to know if it has actually been remapped or not. I suppose that the only way to do so, would be to totally reformat the drive, and see if the bad sector continues to appear.

AFAIK, most disks keep a few spare sectors on each cylinder and at least one spare cylinder. To the extent possible, a bad sector will be mapped to another sector on the same cylinder, or one located close by.

Theres another concept called Sector slipping in which all sectors between the bad sector and the replacement sector move down by one.

 

[Teerex]

I think the drive firmware has control over bad sectors so you can't do anything about it if it's really bad, although I'm not sure. Sectors that were unreadable for some temporary reasons will probably be succesfully read on second pass anyway.

 

[seekermeister]

The reason for my interest is to be able to determine if a bad sector is actually cause for an RMA or not. I got the impression that some file system problems might not be corrected properly by chkdsk, SFC, etc. and would remain until doing a low level format of the drive. Even though that would be a step that I would take prior to an RMA, it is not one that I would want to do, unless convinced that an RMA was necessary. Therefore, it would be good to be able to 100% determine the situation beforehand.

Teerex,

Are you saying that if a sector remains marked bad, after a second run of a diagnostic, that it definitely is a permanent flaw of the drive?

 

[Bill2]

The best route to RMA would be running the concerned diag tool and failing the short and long and other tests. Save a copy of the test results.

Until then, unless the disk is sluggish or there're random read/write errors, it should be fine. As usual, keep your data backed.

 

[Teerex]

For RMA reasons, if the drive is under guarantee, just use a tool that reads SMART. If the value of Off-line uncorrectable sector count is above zero, RMA the drive and use that SMART value as proof that you are entitled for a new drive, even if you haven't lost any data.

Other names for that value are Uncorrectable Sector Count or Offline Uncorrectable or Off-Line Scan Uncorrectable Sector Count.

 

[seekermeister]

That sounds like an idea, but I'm somewhat confused about which value to look at. As can be seen in my screenshot of HD Tune, there are several numbers to check, and all of them are above zero, except the threshold, which I assume is a fixed value. Since all 6 of my hard drives have similar numbers, it's seems unlikely that all are defective in exactly the same fashion.

 

[Teerex]

You have one finally unredeemably bad sector and one pending final evaluation. I say RMA the drive while your guarantee is valid.

 

[seekermeister]

HD Tune just finished a 6+ hour error scan of the same drive that I posted above, and it gave it a green light all the way...no bad blocks. If there is actually a permanently bad sector, why wouldn't it show up on the scan? I do have a couple of other drives which do show a bad block, regardless of how many times that I scan them.

 

[Teerex]

I'm telling you what I see in the SMART readings you posted. Your drive's firmware says that one sector is unreadable - permanently. Was it remapped? I don't know. Try other S.M.A.R.T reading tools, SIW for example. The choice is yours, but from what I'm seeing you have a valid case for RMA. That value should be 0, in fact, I have a 4 year old laptop that still has 0 in that category.

 

[gregrocker]

How does HDTune stack up to the manufacturer's diagnostics/repair extended CD scan, which supposedly can repair bad sectors?
HD Diagnostic
Hard Drive Diagnostics Tools and Utilities (Storage) - TACKtech Corp

I've returned bad HD's to service for years using maker's repair scan and chkdsk /f.

 

[seekermeister]

I'm familiar with most of what is in the two articles that you linked, but one comment caught my eye on the first one:

NOTE: Please run the test on all hard drives attached to your system. This is necessary as a fault in any of the drives can cause problems with the entire system.

Since I do have a couple of drives that I know have some bad sectors, I'm wondering if they could be responsible for some system instability that I have been having, even though neither of them are system drives? In fact, one of them is an external drive, they other one I will have to check, because I can't remember if it is in the PC or HTPC.

Since having installed the new motherboard, CPU & RAM, I haven't been able to boot with either of the two external drives powered up on the PC, it either simply hangs or hangs and starts loudly squawking an alert. I was thinking about replacing the controller card, because the Event Manager contains a series of errors regarding controller errors on drives 5 & 6, which are the two that are external.

 

[gregrocker]

Can you post back a screenshot of your maximized Disk Management drive map with listings, with all drives attached and powered up?

I'd also run Disk Check with /f switch on the externals, if necessary from Command Line as shown in tute.

 

[seekermeister]

Here is the Disk Management screenshot. I have only begun having Windows scan the drives, so that is going to take a while.

 

[gregrocker]

What is on Disk0? Ubuntu? It has no System flag so I assume if it's a bootable OS it is booting via C. If there's no OS on Disk0 I would move the Win7 HD there, making sure it remains set to boot first in BIOS setup.

The WDC 2tb external lettered M is not showing on map, but should not be Active. Where this can be a problem is if it has previously been bootsected and is interfering with boot. I would mark it Inactive using Diskpart from Elevated Command Prompt: Partition - Mark as Active (Method Two)

The Disk Check and definitive HD testing is more important at this point.