WDC Reds Not Suitable?

StephenB wrote:
http://www.cs.virginia.edu/~gurumurthi/papers/acmtos13.pdf finds that disk failures increase as temperature rises. Per Table 3, disks that run at 40C or more fail at a higher rate. Disks running at 50C fail about 1.50 higher rate more often than drives that are < 40C.

Good gosh, how do you find these things? Did you go a-searchin' just for this topic or is it something you already had on hand?

StephenB wrote:
But the downside is potentially shorter drive life.

Bingo. Reds it is then.

StephenB wrote:
There are plenty of ReadyNAS owners using 7200 rpm drives that report temps ~50C.

And all due respect and friendship to them. I'll be taking the road less traveled by, on this one at least.

StephenB wrote:
Increasing the fan speed (not possible in the UI today) would also help, but the study didn't find that to be cost effective in a data center environment (figure 14). Though the NAS used in data centers are quite different from home NAS. Their test servers held 35 drives each, and they were increasing fan speeds from 7000RPM to 12000RPM. That increase only reduced drive temps by 5C.

Sounds like too much trouble and expense for too little gain. Plus wear-and-tear on the fan. (Not a big one, I know, but why enhance the risk?)

StephenB wrote:
InteXX wrote: What would the ReadyNAS Maniac do? I'm using mostly WDC Reds (11 altogether, in 4 NAS), and see temps around 30C in those systems. The performance is good enough for me (generally my systems are either CPU bound or network bound).

Sounds good. Performance isn't a big deal for me--natch, I've got a 104--so I'm confident I'll be OK with the Reds.

StephenB wrote:
I plan to stick with WDC REDS for now. If I start seeing a lot of failures, I'll reassess.

If you come to different conclusions I'd be pleased to know about it, if you happen to think of it at the time.

StephenB wrote:
If I were running a business

Which I am... (one-man dev shop)

StephenB wrote:
(or had the NAS in an environment with conditioned air)

Which I don't... (high on a warm shelf presently, w/plans to move it down toward the cooler floor soon)

StephenB wrote:
I'd probably use RED Pro drives in one or two systems, and see how they worked out.

Which I won't... (for now)

Thanks,
Jeff Bowman
Fairbanks, Alaska

InteXX wrote:
StephenB wrote: http://www.cs.virginia.edu/~gurumurthi/papers/acmtos13.pdf finds that disk failures increase as temperature rises. Per Table 3, disks that run at 40C or more fail at a higher rate. Disks running at 50C fail about 1.50 higher rate more often than drives that are < 40C.

Good gosh, how do you find these things? Did you go a-searchin' just for this topic or is it something you already had on hand?

I recently linked to this article in one of my posts about fan control as well:
viewtopic.php?f=35&t=79560

Quote from the article:
"In summary, we see that average temperature h as a strong correlation to disk failures at different locations in side chassis, rack and across datacenter floor.
However, we do not observe a significant correlation between variations in temperature and disk failures".

So... If I read this correctly, It is more important to keep your drives at a constant temperature than "as-cool-as-possible".
I stand by my argument however that 50c is too higher for a HDD which states a max op temp of 55c.

Also as an interesting side note, it looks like average high temps in Alaska are 23-65F.
I'm living in Singapore where the average high temp is 86-89 all. year. round. ;)

AndreasKa wrote:
86-89 all. year. round. ;)

Sounds like a year-long heat wave to me :D

But the prettier women are there I'm sure of it :)

Thanks,
Jeff Bowman
Fairbanks, Alaska

AndreasKa wrote:
I stand by my argument however that 50c is too higher for a HDD which states a max op temp of 55c.

Very well. Reds it is.

Thanks,
Jeff Bowman
Fairbanks, Alaska

InteXX wrote:

AndreasKa wrote: But the prettier women are there I'm sure of it :) Maybe not prettier, but less clothing worn by everyone... (for better or worse) :D I visited Fairbanks once (many years ago). It was August, but more like spring in Boston (rainy, cool, ...). Alaska's beautiful, but the climate (and the darkness in the winter) is not for me. InteXX wrote: Good gosh, how do you find these things? Did you go a-searchin' just for this topic or is it something you already had on hand? AndreasKa posted it here a few days ago. There's a set of interlocking topics around disks (temps, how spindown affects disk life, bitrot). I like seeing studies on them, and not just random opinions. AndreasKa wrote: I stand by my argument however that 50c is too high for a HDD...

No disagreement from me. I like to see temps in the low 30s (even before I read the paper you posted).

Fan noise is a factor too. At one point I had my duo in the familyroom, and when the fan kicked in during video streaming it was annoying. Some posters have ReadyNAS in their bedrooms. So there is a pick-your-poison aspect to this.

Overall, I look at performance, acoustic noise specs, and expected temps/power use (having learned about acoustic noise the hard way).

StephenB wrote:
I visited Fairbanks once (many years ago).

Well shucky darn! We probably walked right past each other in the hardware store :)

StephenB wrote:
the climate (and the darkness in the winter) is not for me.

That handles the clothing problem ;)

StephenB wrote:
I like seeing studies on them, and not just random opinions.

Good point. I agree.

StephenB wrote:
So there is a pick-your-poison aspect to this.

I've picked the poison in the Red bottle. Got two out of three so far.

Thanks,
Jeff Bowman
Fairbanks, Alaska