[Nottingham] Disc failure in a RAID 1 array

[Andy Smith]

Hi,

On Wed, Sep 26, 2018 at 10:30:35PM +0100, Jason Irwin via Nottingham wrote:
> All SMART tests passed, drive claims to be perfectly healthy with no failures and the RAID is rebuilding.

Does the drive support SCTERC? If unsure, what happens when you do
this?

# smartctl -l scterc,70,70 /dev/sda

(substitute correct drive for /dev/sda)

It's not destructive. More info:

http://strugglers.net/~andy/blog/2015/11/09/linux-software-raid-and-drive-timeouts/

If it gives an error, then your RAID is highly likely to fall apart
the first time there is a read error, and you will need to set the
block layer's timeouts to something like 180 seconds:

# echo 180 > /sys/block/sda/device/timeout

Not supporting SCTERC is a leading cause of drives being kicked out
of Linux RAIDs even though there's nothing much wrong with them.

Cheers,
Andy

-- 
https://bitfolk.com/ -- No-nonsense VPS hosting