"One or more disks are failing" ?

Tony Nelson tonynelson at georgeanelson.com
Mon Jul 6 00:06:47 UTC 2009

On 09-07-04 11:30:52, Scott Beamer wrote:
> Tony Nelson spake thusly:
> I've been getting these warning messages popping up for 2-3 weeks now
> (see original post for details), but I've otherwise not had any
> problems with losing data or the drive's performance.
> I'm not sure how to tell if this number is increasing....

You write down the number and look again later.

> > The OP should enable Automatic Offline Data Collection, which will
> > scan the entire disk for bad sectors "every 4 hours", and remap
> > them if they are still readable.  If sectors are unreadable when
> > that is being done, then/ the disk needs replacement as data will
> > surely be lost (sooner or later).
> >     
> > $ sudo smartctl -o on /dev/sda
> Thanks. I just started that. Does it run indefinitely? Do I stop it 
> at some point?

The disk will do "Offline Data Collection" forever or until you stop 
it.  It can have minor impact on drive performance.  Part of what it 
does is a surface scan "every 4 hours", which will attempt to recover 
any bad sectors, hopefully while they're still readable.

All my disks run with "Offline Data Collection" enabled.  One of them I 
found in a snowbank this January, but all its 300 GiB seem fine.  
Another started giving errors about 7 years ago, but it's been OK that 
whole time.

> I've had the worst luck with hard drives in the past 4 years. It made
> up for the 10 before that with no trouble whatsoever.

Hot drives are a common problem, but not here.  Most drives (and yours) 
show the drive temperature at attribute 194.  Your drive helpfully 
asserts that not only is your drive not running hot, but that it has 
never run hot.  Perhaps it is just luck.

> I just couldn't believe this one might be a goner also.
> I'm crossing my fingers. Meanwhile, since I started this thread I've
> ordered two new drives (one or both of which may replace this one
> and/or supplement it).

I'd suggest setting up a RAID mirror, so that when one of the drives 
fails, you're ready to just keep working.

> Thanks for your help.

You're welcome, and good luck.

TonyN.:'                       <mailto:tonynelson at georgeanelson.com>
      '                              <http://www.georgeanelson.com/>

More information about the fedora-list mailing list