[I have Googled this, but I'm not finding anything that works for me - hence this post].

I'm running Fedora Core 6 on a Dell SC420. Been working fine for ages - on Friday night I did a normal shutdown, but on Monday morning it won't boot - I get a message "Error 16: Inconsistent filesystem structure", at which point it loops round and tries to boot again, with the same error, etc.

If I select a previous version of the kernel at the boot menu, it gets a bit further - it starts booting, but this time runs a "Checking filesystems" - the / filesystem checks OK, but when it gets to /boot I get an error:

"Error reading block 49404 (Attempt to read block from filesystem resulted in short read) while doing inode scan."

It then lets me drop to a maintenance prompt.

If I run fsck from the maintenance prompt, I get the same "short read" error, followed by "ignore error<y>?" Selecting 'y', I get "force rewrite<y>?" - selecting 'y' the fsck then ends.

If I re-run the fsck, it reports the / and /boot filesystems as clean. However, if I reboot, the whole thing starts again.

Can anyone help me? If not, I have a fairly current backup of this drive, so buying a new drive and starting with a fresh install is an option, but I'd obviously rather not.

Thanks in advance for any ideas.

try

fsck -f -y (/dev/xxx whatever)

If you are getting errors on the same blocks each time then it sounds like that drive has hard errors that it is unable to work around. As such you need to be thinking about replacing the disk drive ASAP since this could be an indication that a really bad failure is about to occur. Make immediate backups of all data on that drive, move that data to another disk and use from there.

At the maintenance prompt you can get to, post the /var/log/messages file here. I suspect this may confirm seek errors on the drive.

Thanks for your help - yes, you're right - /var/log/messages is fairly scary looking:

smartd: Device: /dev/hda, 20 Currently unreadable (pending) sectors
smartd: Device: /dev/hda, 20 Offline uncorrectable sectors
[repeats every half hour]

I'm all backed up already, so I think I'll take your advice and go out and find myself a new hard-drive pronto.

Thanks again for the help.

Hi again.

Mmmm. These dont look good do they.

There are lots of tools that can perform surface and low level scans but even if they identified the problem you would still have to fix them (repair the drive).

It could be the controller. Have you tried connecting the drive to another controller (primary<>secondary) NOT master<>slave. Are there any other drives on this same controller?

It could also be temperature of the drive. Is it HOT inside the system case/chassis?

Or age - How old is the disk? Do you know?

Good luck mate anyway.

Hmm, thanks for the ideas. I do have some other similar (nearly identical) machines, so I could do a bit of disk / controller swapping.

The disk is not hot - in fact, it's almost as if it's the opposite - the profile of this machine is that it's on 24x7, and very rarely switched off - when I switched it off last week it was probably the first time for months - and it was when doing a cold (i.e. really cold - the room and machine were cold) restart that this problem occurs. Now it's all warmed up again, I'm tempted to try again. I think the machine is about 3 years old, but if memory serves I think the disk is older than that (long story...)

Realistically, it's an old drive which is running low on space anyway, so I think I'll grab a new one and hope that solves it.

Thanks for your help.