[mythtvnz] Hardware Recommendations

[David Moore]

On 05/10/11 22:20, Stephen Worthington wrote:
> Google's test data pretty much proved that it is a myth that the
> operating temperature of a drive affects its reliability or lifetime.
> They found that as long as a drive was operated within specification
> (<  60 C for most drives now), temperature did not matter.  I still try
> to keep my drives below 50 C anyway, but I am not much bothered by
> high temperatures much any more after I read that report.  The drives
> in my MythTV box are currently reporting 43, 41 and 45 C.  The new
> green 3 Tbyte drive which is external to the box is at 32 C.

I haven't read the actual report yet but did look at a summary here:
http://storagemojo.com/2007/02/19/googles-disk-failure-experience/
Also this article:
http://www.linuxjournal.com/magazine/monitoring-hard-disks-smart
about Smartmontools which claims "Studies have shown that lowering disk 
temperatures by as little as 5°C significantly reduces failure rates, 
though this is less of an issue for the latest generation of fluid-drive 
bearing drives." Maybe the last bit about bearings is the critical thing 
these days. However I'm certain from some experience with MTBF 
calculations and telecom hardware testing and high temp failures that 
higher temps _are_ definitely worse for electronics in general. Having 
said that I don't think there's any need for extra drive cooling unless 
your temps are getting close to the spec limit.

>
> The SMART warnings I usually got from the 7200.11 drives were more and
> more bad sectors being remapped, or pending unable to be remapped yet.
> There were around 1500 spare sectors available for remapping on my
> 7200.11 1 Tbyte drives, and a couple of them reached that limit!
> Usually, one or two bad sectors remapped is not a problem.  But if you
> start to get one a day or more on average, then replace the drive
> quickly as you are likely to have less than 10 days before it is dead.
> And it there is a continuing pattern of more sectors being remapped,
> even if it is not up to averaging one a day, then I would get the
> drive replaced too.
>
Funny thing with my 7 year old laptop. I run Ubuntu normally but booted 
Fedora a few times and it popped up a Smart warning about problems with 
the HDD. So I installed smartmontools on Ubuntu and ran a long test 
which completed without error. There is a warning about "Self-Test Log 
Structure error: invalid SMART checksum." but I don't see this as a 
reason to flag a serious drive problem. So then I started looking at the 
attributes and the normalised values all seem fine. There is a 
"Reallocated_Sector_Ct" attribute and this does have a large raw value 
(262148) so maybe this is what Fedora flagged.

I also ran a long test on my 500 MB 7200.11 (ST3500320AS) and it passed 
with no errors. But the Seek_Error_Rate raw value is 13077544596 (60 
worst normalised with 30 threshold). Only 82 raw value for 
Reallocated_Sector_Ct though. Hmm. Hardware_ECC_Recovered has a large 
raw value (160510852) also.

My two WD Green drives are sitting on zero Reallocated_Sector_Ct and 
Seek_Error_Rate raw values. Models are WD10EADS-00L5B1 and WD20EARS-00MVWB0.

I think I need to do more reading about smart attributes. And maybe back 
up the laptop and the 7200.11. :)