[arch-general] journalctl and I/O errors

[Sébastien Leblanc]

On 8 February 2014 12:06, Janna Martl <janna.martl109 at gmail.com> wrote:
>
> On Thu, Feb 6, 2014 at 1:03 AM, Sébastien Leblanc <leblancsebas at gmail.com>wrote:
>
> > Conclusion (as I understand it):
> >
> > 1. There is definitely a bug in Journalctl: it crashes (segfaults) on I/O
> > errors.
> >
> > 2. You have a drive that is failing, or your BIOS might not be set
> > correctly.
> >
>
> Thanks, all, for the analysis. I have submitted a bug report [1] for
> systemd. Also, it seems you were right about hardware failure --
> though I still can't get smartctl to acknowledge anything being wrong
> (except for nonzero current pending sector count), I ran badblocks,
> which found a bunch of errors, including on sectors corresponding to
> files outside /var/log/journal. Guess I should get a new drive.
>
>
> [1] https://bugs.freedesktop.org/show_bug.cgi?id=74714


Do not rely too much on SMART data; in their hard disk failure study,
Google concluded that when SMART reports a drive as unhealthy, it is
often right, but for many drives that failed, SMART was still
reporting them as healthy.

> "Failure Trends in a Large Disk Drive Population", by Google Inc.
> [...]
> Our analysis identifies several parameters from the drive’s
> self monitoring facility (SMART) that correlate highly with
> failures. Despite this high correlation, we conclude that mod-
> els based on SMART parameters alone are unlikely to be useful
> for predicting individual drive failures
>
> http://static.googleusercontent.com/media/research.google.com/en//archive/disk_failures.pdf

-- 
Sébastien Leblanc