[Info-vax] analyze/disk errors

Stephen Hoffman seaohveh at hoffmanlabs.invalid
Fri Dec 6 18:52:06 EST 2013


On 2013-12-06 22:45:21 +0000, Simon Clubley said:

> I don't know if it's just me, but I think disk failure patterns have 
> changed.  Back in the 1990s, I did see disks with bad sectors 
> developing over time, but these days, I can only think of one disk over 
> the last 5 years or so which developed bad sectors while continuing to 
> stay online.
> 
> Most of the time these days the disks I have experience with either 
> start to fail by dropping offline and then maybe comes back online 
> after a power cycle before dropping off again, or just outright fail 
> with no warning.

I've seen a few that were merrily revectoring sectors until they ran 
out of spares, then the I/O got ugly.
Most of the SMART data was empirically not very useful for predicting 
failures, save for a few data points.
Related: <http://labs.hoffmanlabs.com/node/93>


-- 
Pure Personal Opinion | HoffmanLabs LLC




More information about the Info-vax mailing list