[mythtv-users] Hard Drive reliability esp. RAID issues

[Ian Oliver]

In article <387ee2021003080839r770ea4eew886d296226fcbccd at mail.gmail.com>, John Drescher wrote:
> I would say from experience that this is very unlikely that a second
> drive will fail in the 8 hours it takes to rebuild a 6TB array.

Agreed. The only time I had it was with a combination of mobo + sata card + libata that caused 
errors when everything was stressed. Of course, a rebuild *really* stresses things. Despite doing 
a lot of rebuilds, and having a few fail, I never lost any data. (Yes, I did fix the system ASAP, 
but it was in this dodgy state for a few weeks while I tracked down the issue.

> And
> even if it did with linux software raid you could recover.

As long as the error is transient or localised, then yes.

There are ways to improve handling of bad/pending sectors so that they don't lurk and explode 
during a rebuild. You can read all data during idle time, which allows drives to reallocate any 
sectors that are starting to go bad. The raid software can also cope with unreadable sectors by 
regenerating the data from the good drives and then writing it to the bad disk. Again, this 
allows remapping.

I don't know what of this mdadm does.

Ian