LinuxQuestions.org
Help answer threads with 0 replies.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Mandriva
User Name
Password
Mandriva This Forum is for the discussion of Mandriva (Mandrake) Linux.

Notices


Reply
  Search this Thread
Old 09-21-2005, 02:55 AM   #1
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Rep: Reputation: 16
Unhappy Hard drive; file system error; fsck failure


I wasn't there when it happened, but apparently my wife was watching the end of a movie on the DVD player when the hard drive started making some reasonably loud noise. Having finished the movie she shutdown the system. According to her that took a long time. The next day I get all of this:

On boot:

/dev/hda6 contains a file system with errors, check forced

/dev/hda6: Unexpected Inconsistency; RUN fsck manually. Failed to check filesystem. Do you want to repair the errors?

I say yes. And then, after many minutes, I invariably get the following response, though always with different block numbers, sometimes higher, sometimes lower:

Error reading block 7143619 (Attempt to read block from filesystem resulted in short read) while doing inode scan. EXT3-fs error (device hda1): ext3_find_entry: reading directory # 32705 offset 0.

When I answer 'no' to Want to repair errors? I get dropped to a shell:
(Repair filesystem) 1#

I tried many times to boot normally, tp boot in failsafe; no matter. Finallly I tried running my Installation CD, this is what happened:

Failed to check filesystem /dev/hda6 repair errors? [yes]

...minutes later...

fsck failed on /dev/hda6 with exit code 4 or signal 0


What the helll is going on?! I'm running a Dell GX150, 1Ghz on Mandrake 10.1
 
Old 09-21-2005, 10:58 AM   #2
Maldain
Member
 
Registered: Jan 2005
Posts: 112

Rep: Reputation: 15
Ouch. I suspect if you open your case and look you'll see that the jumper on your hard disk is set to let the cableing decide if your hard drive is a master or slave. For some reason every once in a while that really screws up linux. The drive needs to be either a master or a slave both on the drive and in the bios of your computer. That's the good news the bad news is that your installation is probably toast. You'll probably have to reformat and re-install. Having been there on that one I feel for you. Just thank your lucky stars it wasn't a mission critical server (how I learned my lesson).

M. Lacy
Western Tool Supply
 
Old 09-21-2005, 02:01 PM   #3
Bruce Hill
HCL Maintainer
 
Registered: Jun 2003
Location: McCalla, AL, USA
Distribution: Arch, Gentoo
Posts: 6,940

Rep: Reputation: 129Reputation: 129
I would check the drive for errors. You can do one of two things.
Navigate to the website of the drive manufacturer, and download
their diagnostic utilities. Usually this can be for either a floppy or
a CD. Make the appropriate medium, boot the computer with it,
and do a thorough check of the drive. In Linux you can also boot
with a LiveCD (perhaps your distros CD .... don't know diddly about
Mandrake) and run "e2fsck <blah>" on it. You'll need to issue and
read "man e2fsck" of use Google <Linux> to get more information
on what options, and how to journal that partition first. Sorry I don't
want to give them to your incorrectly. I've never run any Linux part
except ReiserFS, so I've no experience with ext3.
 
Old 09-22-2005, 02:04 AM   #4
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
[QUOTE]Originally posted by Maldain
[B]Ouch. I suspect if you open your case and look you'll see that the jumper on your hard disk is set to let the cableing decide if your hard drive is a master or slave. For some reason every once in a while that really screws up linux. The drive needs to be either a master or a slave both on the drive and in the bios of your computer.

Assuming that no one else posts something constructive, how do I avoid this problem if I reinstall?

-liNUBEx
 
Old 09-22-2005, 02:06 AM   #5
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
Quote:
Originally posted by Chinaman
I would check the drive for errors. You can do one of two things.
Navigate to the website of the drive manufacturer, and download
their diagnostic utilities. Usually this can be for either a floppy or
a CD. Make the appropriate medium, boot the computer with it,
and do a thorough check of the drive. In Linux you can also boot
with a LiveCD (perhaps your distros CD .... don't know diddly about
Mandrake) and run "e2fsck <blah>" on it. You'll need to issue and
read "man e2fsck" of use Google <Linux> to get more information
on what options, and how to journal that partition first. Sorry I don't
want to give them to your incorrectly. I've never run any Linux part
except ReiserFS, so I've no experience with ext3.

I went to the Dell website, but nothing helpful there. I also tried using my distro CD, but that was no help. Is e2fsck different from fsck? What does it mean to journal a partition.

Thanks,
liNUBEx
 
Old 09-22-2005, 07:24 AM   #6
Bruce Hill
HCL Maintainer
 
Registered: Jun 2003
Location: McCalla, AL, USA
Distribution: Arch, Gentoo
Posts: 6,940

Rep: Reputation: 129Reputation: 129
Quote:
Originally posted by linubex
I went to the Dell website, but nothing helpful there. I also tried using my distro CD, but that was no help. Is e2fsck different from fsck? What does it mean to journal a partition.

Thanks,
liNUBEx
I understand that Dell doesn't provide much, if any, computer specs.
(Something to remember before your next computer purchase IMO.)

You can run as root "hdparm -I /dev/hda" to determine your drive,
or "cat /proc/ide/ide0/hda/model". If that only gives you something
like IC25N060ATMR04-0 just enter it in Google and you'll find out
the manufacturer. If you can't determine who made your drive from
the hdparm output, take the side of the case off and look on the drive.

You will need to use e2fsck for that ext3 extension. Notice in my previous
post I stated
Quote:
"You'll need to issue and read "man e2fsck" of use Google <Linux>
to get more information on what options ..."
There are some differences in the information contained in the man pages
for "fsck" and "e2fsck" which you should read. I think however, in your case,
it would be much simpler to get the manufacturer's diagnostic utility and run
that program.
 
Old 09-23-2005, 10:17 AM   #7
keogk
LQ Newbie
 
Registered: Sep 2005
Posts: 21

Rep: Reputation: 15
Do you still have the dell utilty partition on the drive. You can try to hit F10 or F12 on startup and see if you have a option for HD diagnostic.
Also you can try and run Seagates HD diagnostic on the drive. It is one of the few that will work with drives from other vendors.
I used it yesterday on a Western Digital drive.
 
Old 09-29-2005, 01:56 AM   #8
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
Ok, so I downloaded the disk utility from IBM. I ran all the tests I could, including the 'exerciser' for three hours - but the drive seems to be fine. I did try another boot of the system, and this is what I received:

respawning too fast, journal has aborted
cannot execute /sbin/mingetty

I'm going to try the Seagate HD diagnostics next and see if that helps. If this doesn't work though, is there a way I can save some of my files on the drive? How do I avoid this problem?

-lin
 
Old 09-29-2005, 07:10 PM   #9
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
I tried the Seagate utility and this was returned:

SMART Test: Failed

The following sectors may be damaged:

LBA: 8356444
Status: Write-reallocation SKIPPED
Usage: Not resolved

I received the same message for LBA: 8356444 - 8356449

What troubles me even more is that under the File Structure Test Result, I get this:

Partition 1 (Unknown (6.2 GB) No Name) Unsupported or Unknown Partition

Partition 2 (Unknown (1.1 GB) No Name) Unsupported or Unknown Partition

Partition 3 (Unknown (32.5 GB) No Name) Unsupported or Unknown Partition

Am I screwed?
 
Old 09-30-2005, 07:12 AM   #10
springshades
Member
 
Registered: Nov 2004
Location: Near Lansing, MI , USA
Distribution: Mainly just Mandriva these days.
Posts: 317

Rep: Reputation: 30
Have you opened up the case to look at the harddrive yet? You really should run the diagnostic for YOUR brand of harddrive, and the manufacturer should be marked on the harddrive and quite easy to see. Dell didn't make your harddrive, so they aren't the place to look. They didn't make your processor or your motherboard. They simply buy these things from other places, put the computer together, then sell it as a complete product. You'll find the best information about your harddrive from the company that made it. Most likely it will be something like Western Digital, Seagate, Maxtor, or Hitachi. Taking the side off your computer is pretty easy and may very well be the fastest way to tell considering the state that your computer is currently in.

Those messages from the Seagate utility COULD mean your hard drive is fried. However, they could also mean that for some reason, the Seagate utility just couldn't communicate properly with your drive (which could be the case if it's a brand that their utility doesn't support). Also, the utility for YOUR brand of harddrive may be more adept at fixing any errors it comes across, possibly saving your files and data.

My final thought. If you DO find out that your hard drive is fried, one thing that seems to work for some people is to throw the hard drive in a freezer over night. If this DOES work, it's supposed to be a very temporary thing and so if you try it, make sure everything you need to back up your files is on hand as the hard drive may only be accessible for a very short period of time. If it works, it's much cheaper than data recovery fees.

Last edited by springshades; 09-30-2005 at 07:13 AM.
 
Old 09-30-2005, 06:25 PM   #11
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
Hey Springshades,
I did look at the drive a few posts ago, but on your advice I went back to the same utility, ran it again, and this time it worked (I think I loaded the wrong portion the first time through). Regardless, I have my system again! However, I'm here long enough to save my data and then install the system again, or hold out for Mandriva 2006. This is what I've been planning to do for months now as it is - this fiasco just moves the timeline faster once it is released.
I do have another question though: can I reformat this drive? Or should I just buy another one? And what could I have done in the first place to have avoided these bad sectors?

Thanks Springshades.. and others
 
Old 09-30-2005, 08:14 PM   #12
springshades
Member
 
Registered: Nov 2004
Location: Near Lansing, MI , USA
Distribution: Mainly just Mandriva these days.
Posts: 317

Rep: Reputation: 30
Quote:
I did look at the drive a few posts ago, but on your advice I went back to the same utility, ran it again, and this time it worked (I think I loaded the wrong portion the first time through). Regardless, I have my system again! However, I'm here long enough to save my data and then install the system again, or hold out for Mandriva 2006. This is what I've been planning to do for months now as it is - this fiasco just moves the timeline faster once it is released.
Haha, this is totally my fault. For some reason I forgot that IBM made hard drives. Guess you learn (or remember) something new every day.

Quote:
I do have another question though: can I reformat this drive? Or should I just buy another one? And what could I have done in the first place to have avoided these bad sectors?
I don't think there is a short answer to this. One harddrive that I had failed every couple months with bad clusters til it finally bit the dust for good. It just wasn't a good drive. I'm sure that it's also possible that your harddrive is in perfect condition now that those clusters are fixed. I have a stick of RAM with bad addresses that used to freeze windows every couple hours. Now with the badram patch in Linux (which simply ignores the bad addresses), it's worked perfect for almost a year now. You could have either experience.

You absolutely should be able to reformat the drive and use it for now. Reliability is the main concern. If money is more important to you and you don't have very vital files, stick with this hard drive. If 50-100 dollars (depending on the drive it could be more) isn't too big of a deal and it's going to save you the possibility of another headache like this happening, it may very well be worth getting a new drive. You could then maybe use your current harddrive as extra space for files that aren't as important.

As far as what to avoid... shutting down the computer improperly can mess up your file system, but it usually won't cause bad sectors on the drive itself. Does your computer get extremely hot inside on a normal basis? Heat is always something that can cause issues, but it's pretty unlikely that it's your issue unless you don't have any ventilation in your case. Most likely, there isn't anything you could have done to prevent it. It was probably just a hardware issue of some sort. I know that isn't comforting news though.

Good luck with your upgrade.
 
Old 10-02-2005, 06:51 PM   #13
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
Ugh. After a 45 minute routine that I half to go through everytime, I can gain access to one of four accounts though GUI. I burned the data on that one account, but I can't get access to the others through a GUI to burn those files. I do not have access to my root GUI user. I have tried changing the security level and other GUI-able tricks to gain access to the other files, but no such luck. I do however have access to those directories through a console as su in my one working account. I have roamed around with the command line to see that the needed files are still there, I just can't burn any of them using the command line.
I assume that I should be able to move those files (folders) to the working GUI account using some console commands, but I don't know how. Once I get them to the one workign GUI, then I can burn them like the others files. How then do I move those folders (there are a lot of files to do so one by one)? Or can I change the permissions on that home directory to gain access from mu one useable account? Again, I know I am unable to do so using any GUI command.


-Thanks
 
Old 10-03-2005, 03:40 AM   #14
springshades
Member
 
Registered: Nov 2004
Location: Near Lansing, MI , USA
Distribution: Mainly just Mandriva these days.
Posts: 317

Rep: Reputation: 30
There are several ways to do this. It depends on how much stuff works. Can you su to root while you're still in GUI? In KDE or Gnome or whatever you use? If so, you can su in the terminal, and then put in k3b (assuming you use k3b to burn CDs, if you use another program, you can use that command instead). This will run your CD burning program as root user, so it will have full priviledges to all files. The other thing you could do is:

chmod -R 777 /

though I've never done that before and I'm not quite sure what it will do. It's probably safe to do:

chmod -R 777 /home

that will give you full access to all of the files in any of the user's directories as long as that's all you need. chmod changes permission, 777 gives full read, write, and execute priviledges to everyone, and the -R makes it apply to every folder and file contained within the folder you're using it on. Again, the first version is kind of scary because it gives everyone full access to every file on your computer. I'm not even completely sure if it would work. You'll have to do the chmod command as root.

You can do it either way.
 
Old 10-03-2005, 07:29 PM   #15
linubex
Member
 
Registered: Mar 2005
Location: California
Distribution: mandrake 2008
Posts: 64

Original Poster
Rep: Reputation: 16
Ok, so I used 'chmod -R 777' and then waited a couple of minutes for it to run through ALL of my files - but it worked. I was able to burn all the files I wanted. Interestingly, after about 20 minutes, I mysteriously lost access to those files again. However, this time I ran chmod -R 777 /home and got access in seconds. Weird. Oh, and when I typed GUI, perhaps I just should have typed gui (graphical user interface). As it is, I'm a Gnome guy.
I tried to format the hard drive with my hitatchi disk utility, but after many minutes I got an error along the lines of equipment malfunction: time to buy a new hard drive. I think I'll install Mandriva 2006. Thanks for all your help SS.

-LiNUBEx (but less so now)
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
Hard drive file system problem, I think...please HELP QUICKLY rivalmist Fedora 7 01-23-2009 08:26 AM
Reiser file system / Hard Disk/ Hard Drive Problems Oxyacetylene Linux - Software 4 10-10-2005 02:24 PM
Hard drive; file system error;automate fsck jumbo25 Linux - Software 1 10-05-2005 12:08 AM
fsck and lost hard drive space LinuxGeek Linux - Software 1 07-19-2005 08:17 PM
Hard drive failure error? (sense key Medium error) Arodef Linux - Hardware 2 11-02-2004 12:39 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Mandriva

All times are GMT -5. The time now is 06:11 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration