constant errors file system

[mrajdl]

I don't know why, I don't use the cpu that much, but every day I have to reboot to recovery mode, go to root shell and run fsck. It fixes a bunch of errors, then every thing is fine for a day or 10 minutes. I'm a idiot when it's linux don't know what to do?

Mike

[mrajdl]

Had to do so the same this morning to get it running. It acts like the power is being cut off, but its not cause I always shutdown. When I get it running and launch Chromium it says didn't close properly, and gives an option to open tabs??? That's a bunch of bull I always exit my browser and shutdown or suspend the system.

Ubuntu 13.04, however it started with 12.10. In 12.10 symptoms were the same, but errors were a little different it would tell me the drive for /tmp was not ready or not present.

[mcduck]

Check the SMART status of the hard drive with the Disks tool, the errors could be caused by the drive being close to failing.

...also knowing the exact error you are getting would be helpful.

[mrajdl]

Check the SMART status of the hard drive with the Disks tool, the errors could be caused by the drive being close to failing.

...also knowing the exact error you are getting would be helpful.

My first thought was the hard drive failing, but I never get any errors to hardware failure. Not sure how to check SMART status of hd with Disks tool, does that need to be installed?

Tried to get the errors when at root, with "fsck > errors.txt" , but says root is read only.


Been married to a red head for 13 years! Don't mess with the red or the root!!!

[mcduck]

Disks is installed by default, just search for it in the Dash.

Once opened, select your hard drive, and then in the upper right corner you'll find a small button (with a gear icon), click that and select "SMART Data and Tests".

[mrajdl]

Disks is installed by default, just search for it in the Dash.

Once opened, select your hard drive, and then in the upper right corner you'll find a small button (with a gear icon), click that and select "SMART Data and Tests".

Ok did that, I could click every option but the SMART Data and tests.

I'm to starting think I should just to low level format the hard drive, and start fresh.

[mrajdl]

Disks is installed by default, just search for it in the Dash.

Once opened, select your hard drive, and then in the upper right corner you'll find a small button (with a gear icon), click that and select "SMART Data and Tests".

Ok figured out the SMART stuff it was not enabled in the BIOS.

SMART Data and Tests tells me that there are 763 sectors failing. 763 in a 2 gig drive is not that many. the hdd is close to 4 yrs old.

Are there utils I could run to map out those sectors, or should I just start shopping for a new hdd, oh is it backed up, of course not, and I'm out of blank CD's stupid idiot.

[binaryhermit]

Have they even made 2 GB hard drives in the last 4 years? Heck, the company my mom used to work for was deploying computers with 4 GB hard drives in 1998.

[efflandt]

Usually the symptoms of a bad drive are that you (even with sudo) do not have permission to write anything to the hard drive because, as redmk2 says, it automatically remounts read-only. One problem with that is that there is no permanent log that anything went wrong (and could be why apps think they have not shut down properly). The only way to check for such errors is to look at dmesg when something strange happens "before" shutting the computer down.

My Linux partition is at the far end of a 1 TB drive, which seems to have started going bad months ago when I started having write permission (read-only) issues. Although, e2fsck with the switch to mark out badblocks from Ubuntu on another drive (SSD) helped for a few months, eventually it started getting more errors. So I used Acronis to image Win7 partitions (which had not gone bad yet) and recreated them on a new drive. And I reinstalled Ubuntu 12.04.2 and copied my old home partition to that.

A hard drive reserves some space to transparently replace (remap) good sectors in place of bad sectors. But once all the error sites are used up, it is never going to get any better, just worse. And it should not be trusted for anything essential.

[mrajdl]

Usually the symptoms of a bad drive are that you (even with sudo) do not have permission to write anything to the hard drive because, as redmk2 says, it automatically remounts read-only. One problem with that is that there is no permanent log that anything went wrong (and could be why apps think they have not shut down properly). The only way to check for such errors is to look at dmesg when something strange happens "before" shutting the computer down.

My Linux partition is at the far end of a 1 TB drive, which seems to have started going bad months ago when I started having write permission (read-only) issues. Although, e2fsck with the switch to mark out badblocks from Ubuntu on another drive (SSD) helped for a few months, eventually it started getting more errors. So I used Acronis to image Win7 partitions (which had not gone bad yet) and recreated them on a new drive. And I reinstalled Ubuntu 12.04.2 and copied my old home partition to that.

A hard drive reserves some space to transparently replace (remap) good sectors in place of bad sectors. But once all the error sites are used up, it is never going to get any better, just worse. And it should not be trusted for anything essential.

Ok my hard drive has is going south, thanks for who helped redmk2d, efflandt, and mcduck