5

After 222 days of usage my external usb drive's S.M.A.R.T. system warned about having too many bad sectors on the disk (65537 reallocated sectors).

I tested it with badblocks -svw and no bad block were found. I re-partitioned and re-formatted drive. It seems to work just fine, but my Fedora is still complaining about the drive is dying and I that yellow exclamation sing is always shown on the notification panel.

I'm a bit confused. So here's the questions:

  1. If SMART is right and the disk contains so many bad sectors, why badblocks couldn't find any?
  2. How many spare sectors are there? As far as I know while the bad sectors are successfully remapped there is no reasons to worry, no?
  3. I agree to have less space for the data, but reserve more for the spare zone. Is it possible to negotiate this thing with SMART and calm it down until it's out of the spare sectors again?

Thanks.

|improve this question

migrated from serverfault.com Dec 5 '10 at 20:54

This question came from our site for system and network administrators.

5

Yes, it is.

I tested it with badblocks -svw and no bad block were found

Reality check. Smart checks physical sectors. Remapped sectors are available again to the OS. Basically, you have 65537 secotrs that dioes and now live from the reserve, but your OS is not able to see them. This is why they were REALLOCATED.

How many spares there are is not determined. The disc manufacturer tells you.

Smart is not negotiable normally. it is done by the firmware. That said, what you ask for is not sensible either.Discs do not fail linearly. They wear out. They get faster killing items to the end. So, a "larger reallocatoin space" would pretty much be useless. This is like saying "hey, my tire profile is too shallow, maybe I can just drive longer". Sur you can ignore all warnings, but at the end your tire is used up.

65537 discs are way too much for me to be comfortable with.

Given 222 days of usage, you are within warranty. Get a replacement.

|improve this answer
  • Couldn't it be a firmware problem? I'm just curious why 65537 (2^16+1)? Is it a coincidence? (I don't have papers for this dive, so the warranty won't work for me) – wajda Dec 5 '10 at 20:44
  • They set the warning threshold so obviously it is going to be a round (in binary) number. – JamesRyan Dec 5 '10 at 22:47
  • The threshold seems unlikely to be so high... I would get a second opinion from some different software. On Windows, SpeedFan and HD Sentinel are great, I don't know what SMART tools are available on Linux. – ZoFreX Dec 6 '10 at 0:51
  • 3
    I'm guessing that you have only 1 or 2 reallocated sectors. I'm seeing a similar situation on one of my Hitachi drivs, where Fedora 15 claims I have 65537 bad sectors. However, that is its interpretation of the Raw value from the drive. Which in my case shows 0x010001000000. That is 01 00 01 00 00 00. The encoding of these fields are up to the manufacturer, but I'm guessing my drive only has 1 bad sector, and the other 01 byte means something else. – Jonathon Reinhart Jan 24 '12 at 16:37
6
  1. badblocks can't find bad blocks because they are hidden by the disks' firmware.
  2. If a disk has so many bad sectors in such a short time, there is something seriously wrong.
  3. No

I wouldn't trust this disk a second more.

|improve this answer
1

A high number ob bad blocks is a warning sign for imminent total failure. Backup your data immediately!

|improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.