Re: Recent drive errors

[Thomas Fjellstrom <thomas@fjellstrom.ca>]

On Fri 22 May 2015 03:38:06 PM Mikael Abrahamsson wrote:
> On Thu, 21 May 2015, Thomas Fjellstrom wrote:
> >> for x in /sys/block/sd[a-z] ; do
> >> 
> >>          echo 180  > $x/device/timeout
> >> 
> >> done
> > 
> > I presume it's ok to do that even if the drives do ERC/TLER? Just woke up,
> > but my brain seems to be telling me it shouldn't break anything since the
> > ERC drives should always return after 7s no matter what...
> 
> Correct, the only downside is that if the drive really dies, it's going to
> take longer to detect this.
> 
> I'd rather have longer timeouts to make sure drives never kicked out
> because of lack of ERC or something else, than to have drives kicked for
> some reason (controller reset, lack of ERC, or something else).
> 
> I really really want to avoid drives being kicked for any other reason
> than them really being dead. I'd rather have reads stalled for a few
> minutes than this happening. For other use-cases, the requirements may be
> different.

Yeah, I agree. Especially given this is just my home nas and nas backup setup, 
It's better to just stall rather than die horribly.

-- 
Thomas Fjellstrom
thomas@fjellstrom.ca