WD Red safe operating temperature?

[ghackett]

There are lots of different opinions on the web on this topic (including what I've been going by here: http://www.buildcomputers.net/hdd-temperature.html), but I wanted to get the FreeNAS experts' opinions on safe HD operating temps, specifically for WD Reds since they seem to have become pretty popular.

I'd like to break it down into 3 questions...

1) What is the optimal temp range I should be aiming for to maximize drive lifespan?

2) At what temp should I start to worry, and look into more/other cooling options?

3) At what temp should I shut off the machine (or disconnect the drive, if only one) until the issue is resolved?

Personally, I've been seeing my 8x4TB Reds hovering in the mid-high 30's during the day and low 30's at night (when the heats off). However during heavy loading I've seen it peak around 40-42C (although usually just on one or two drives). I did think this was safe until I started reading some threads where some folks were extremely adamant about never letting a drive go above 40C, so I thought it would be worthwhile to some more definitive opinions on the subject.

 

[cyberjock]

1. Optimal temp range is 30-40C

2. I have my SMART informational temp at 39C, and I make sure I never break that.

3. That's a personal preference, but I try my darnest to stay below that 39C. Once you go above 40C things head downhill, and at 45C they get significantly worse. I'm pretty adamant about not breaking that 39C rule. I don't have central A/C so my server hides in the basement in the summer. Hard drive temps do sometimes go above 39C but I make sure to schedule scrubs only on cooler days of the week, etc. Basically I baby the heck out of my drives and I've enjoyed very low failure rates.

Read the Google white paper on hard drive life and you'll see the cool chart that shows hard drives between 40-45C see a rapid rise in failure rates. Its somewhere around double the 35-40C failure rate, which 45-50C even higher.

http://static.googleusercontent.com/media/research.google.com/en/us/archive/disk_failures.pdf

 

[Fox]

1. Optimal temp range is 30-40C

I just checked the google study document (pdf), and it looks like to me the optimal range is 35-45C.

Here is a quote from the actual study (paper):

The figure shows that failures do not increase when the average temperature increases. In fact, there is a clear trend showing that lower temperatures are associated with higher failure rates. Only at very high temperatures is there a slight reversal of this trend.

It seems the average failure rate doubles, like you said, though above 45C, but it only goes from 1% to 2% (on average). On the other hand, drives with temps lower than 35C, actually see higher failure rates as you go below 35C, and much higher failure rates below 30C. Here is the graph from the google doc:

Read the Google white paper on hard drive life and you'll see the cool chart that shows hard drives between 40-45C see a rapid rise in failure rates. Its somewhere around double the 35-40C failure rate, which 45-50C even higher.

I'm not sure I agree, but the bigger problem is that the OP said his drives drop down to the low 30's which means he might be at worse risk with the cooler temps.

I am also equally concerned, my drives are currently running at or below 30C.. I think there is one at 28C.. I think I need to remove some fans..

 

[panz]

I don't know if those charts are going to take into consideration things like "cross factors", for example interactions between temperatures and vibrations.

My personal (little, limited and "non professional") experience with my last server (a machine running 24/7 with a mixed-brand of Seagate, Maxtor and WD) that lasted 14 years: I discovered that hard drives were literally boiling, with temperatures ranging from a minimum of 45° C to a maximum of 72° C!

One thing I have to say is - I'm sure of this - more damage to hard drives is caused by "dirty" power = not having a good UPS.

When I first ran that server I hadn't an UPS and hard disks were failing a lot. After installing the UPS all problems went away.

Just my 2 cents.

 

[cyberjock]

Fox- Read the whole document... That's only one chart for temperature.

 

[Fox]

Fox- Read the whole document... That's only one chart for temperature.

I have read it.. There was only one other chart that showed an increase in failure rates for higher temps in the third+ years, but in fact many of the graphs show that the third year is just not the best for drives that are either hot or have a high utilization. And of course, heat could be a result of utilization (duty cycle) vs just being in a hot case or more likely a combination of both. Although I show the graphs, the statements in the article seem to favor not obsessing over temperature. Of course everyone should form their own opinion. I think the study was good, but not truly definitive as far as drive temperature is concerned.

Also, I noticed they gathered the data for the temperature study over a 9 month period. Not over the full life of the drives. So really, I wonder if that type of snapshot is really valid. I would rather see the life of the drives when they are at a given stable temp. Of course, that data would be much harder to gather.

Here is the second chart:

Here is another good article on the topic (which in turn has other good links as well)

https://storagemojo.com/2007/02/19/googles-disk-failure-experience/

Some quotes from the Google report:

Surprisingly, we found that temperature and activity levels were much less correlated with drive failures than previously reported.

Contrary to previously reported results, we found very little correlation between failure rates and either elevated temperature or activity levels.

The figure shows that failures do not increase when the average temperature increases. In fact, there is a clear trend showing that lower temperatures are associated with higher failure rates. Only at very high temperatures is there a slight reversal of this trend.

What stands out are the 3 and 4-year old drives, where the trend for higher failures with higher temperature is much more constant and also more pronounced.

 

[D4nthr4x]

So basically right before your drives turn 3 years old just turn them off for a year and they'll be good :D

 

[Whattteva]

One thing I have to say is - I'm sure of this - more damage to hard drives is caused by "dirty" power = not having a good UPS.

+1 to this. I had a machine that had a hard drive failure cause of a power outage. They had backup power, but it doesn't kick in until 5 minutes after the main line power outage.

 

[SirMaster]

I've also seen a trend with very new hard drives in the last year or so ship with much higher manufacturer rated operating ranges.

I'm not sure why the increase, but a few years ago most hard disks were rated for a maximum of 50-55c, but that has increased to 60c (current seagates) and 70c (WD RED) more recently.

Obviously these are very hot temperatures and you would have to have probably have no fans to get anywhere close. I wonder if the new hard disks do anything to the failure rates at temperatures in the 45-50c range though. None of these studies really say which drive models were tested and the studies are now several years old so the information may not be 100% relevant anymore.

Also what is going to put even more stress on a hard disk is the fluctuation in temperature of cooling and heating. This of course is going to expand and contract the parts inside the hard disk and that can really start to cause things to break down over time.

 

[Whattteva]

Not sure of the applicability of this to HDD's, but from a semiconductor standpoint, you could usually do this through binning.
AMD did this with their R9 290's that are rated to run at 90 C all the time. As manufacturing process matures (they've been using the same process for over 2 years I believe), your components can tolerate higher temperatures and still perform reliably.