fsck error on boot (Ubuntu 9.04)?

Sometime over the past month or so, I started getting an error from the scheduled disk check that runs every so many boots. I've attached a picture of the readout.



When I try running a manual fsck, I end up with basically the same problem this guy has, except more block write errors. Nothing seems to get repaired. And then the process repeats every few boots.

As a result, I decided to run HDTune, and here are screenshots of my results.









As you can see, the Health is supposedly OK, but there is a bad sector in the scan. Could that be the cause of this Ubuntu problem, or is it maybe a corrupt installation instead?

If it is caused by that disk corruption, should I run e2fsck -c to mark that as bad and not use it, or what?

TIA,
-pixelot

 

I had this error a while back and I think it happened because there was an error in /etc/fstab that caused /boot to point to the wrong partition. I believe that to be the case because I had copied over an /etc/fstab from an old install without updating the UIDs. In any case, I decided to just reinstall the OS since it was new anyways.

 

Sorry to tell you this but a bad sector is actually a bad part of the hard drive. That is time to replace the hard drive. Right now I only see one but in time it will start multiplying. The good thing is that your OS knows where it is and it wont write information because it will get lost if it does. Time to see if you laptop has any warranty, if it does, they will usually sent you a replacement.

 

Well, thanks for the info you guys.

I may reinstall, or I may just sit it out, in which case should I go ahead and run e2fsck -c?

At any rate, I was planning on upgrading my laptop HDD to something bigger, and maybe the RAM and wireless card too (if that's possible).

So if my HDD is dying, it's not a huge deal. About the warranty... I'm not sure if I'm still under warranty, but I think I am. So I guess I could go ahead with that idea. I wonder if Dell voids the warranty if you decide to upgrade your own? If that's allowed, I could

1. Use the one they send me for an external
2. Upgrade my laptop HDD
3. ???
4. Profit!!!

Of course if my warranty is up soon anyway, I could get a replacement HDD from them, and then do whatever upgrades I wanted and just sail off into the brave world of warranty-free waters.

 

Pix, I would replace the internal HDD, and maybe use the one you take out as an external and keep and eye on it. Back when I used to keep desktops forever, I'd wear out HD's and the bad sectors would proliferate over time at which time I'd swap the drive.

I might keep my /home and do another install and see what happens....with 80gigs, you need another HD anyway....

 

Boot from a LiveCD/Live USB, run gparted and do a check. It will output a log file you can study.

how old is the HDD?

 

Replacing the HDD because of one bad sector is pointless, wasteful, and impractical unless you see the problem worsening. As long as you keep up-to-date back ups of the machine in case of a total failure, there is no reason to replace an otherwise perfect hard disk.

There are built-in utilities in most Linux distros that allow you to add bad sectors to a black list for future reference. See here: http://linuxgazette.net/issue42/tag/10.html

 

I basically agree with you, but I also agree with everyone else.

I've got five year old harddrives with 2-3 bad sectors that work just fine, but I have also had drives where the first bad sector is a warning sign of a bad sector cascade into storage hell.

One never knows, that's why one keeps back ups.

The key problem with a drive with a bad sector is that most of the time, you cannot image it. That's why I typically replace a system drive when it develops bad sectors and relegate the drive with a bad sector to simple storage (which I back up file by file, rather than by image).

 

Thanks for the input... I do keep my info current on three HDDs at all times, so if it fails, I'm not in the proverbial soup.

However, I would like a bigger HDD, so maybe I'll get a bigger one, and then use this as an external, or maybe stick it in my mom's laptop, which has eaten the last two drives I put in it. This time I might try using a cooling pad or something, in case it is a heat damage issue, as was suggested in my thread about it.

Ah, thanks. That's the e2fsck -c command that I'd read about.

 

I'd first find out if it was indeed a bad sector, or some file corruption.

 

Hmmm. Would HDTune show it as a red block if it was just file corruption?

 

Back up your data ASAP if you haven't already -- you can use gddrescue on a live CD. You can then take your drive image and fsck it with another system.

Check the logs for I/O errors after you rescue the drive image - it's possible your drive is failing but it could be the controller failing or even a software bug.

Edit: https://help.ubuntu.com/community/DataRecovery#Data%20Recovery%20from%20damaged%20filesystem%20or%20drive

 

Just as an update... I did end up running e2fsck -c because the drive really started acting up, and it actually seems to have helped for now.

After doing that, I was able to run HDTune, which came up clean. So I think I'll be able to get along until I get a new drive. Thanks again for the help!