I have two new-ish 1TB hard drives in my server, both running at 38 degrees C at the moment. Should I be worried?

  • Hi, I wasn't logged in when I posted the OP so can't comment on your answers directly. @Cakemox, thanks for that, seems like I have no worries :) @Iain, the specs say 0-55 degrees C, however I heard you should take spec sheets with a pinch of salt. Adding to that, according to Google's research (which I've just been looking at), SMART data and manufacturer ratings aren't a very good indicator of imminent drive failure. @Paul, that sounds like a bug in the SMART reading software. Have you done a smartctl -a on the drive? – Fela Maslen Apr 9 '11 at 13:56

38C should be fine. A quick look at some spec sheets shows that operating temperature is 0C to 60C on many models.

You might be interested in section 3.4 of a study done by Google on hard drive failure rates. Temperature is not an issue in drives until they are a few years older, and even then, it seems, it's not much of an issue under 40C.

| improve this answer | |

What does the manufacturer's datasheet/documentation have to say about temperature ?

| improve this answer | |

Be careful of the interpretation of your sensors. I have Munin running, and the smartd results shows my drives have "Temperature_Celcius" of around 230.0 (which if true would probably indicate they were on fire) but another probe called "HDD Temperature" records them around 27 and 29 C, which seems much more likely to me.

| improve this answer | |
  • 2
    Most likely the 230.0 should be 23.0, some sensors provides results in 1/10th of a degree, which a lot of programs interpret wrong. – bjarkef Apr 9 '11 at 16:46
  • I think 230.0 is 'VALUE' of smartctl outpit, not 'RAW_VALUE'. Value shows how far your drive from being fault. For example, this is output from smartctl on my note: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 194 Temperature_Celsius 0x0022 091 085 000 Old_age Always - 56 It's templerature is 56 (from raw_value column), not 91 from value col. – rvs Apr 10 '11 at 14:35

Slightly OT: If you're concerned, I find hddtemp on Linux indispensable:

# hddtemp /dev/sdb

/dev/sdb: ST9160412AS: 43°C

You can run it as a daemon or script around it very easily.

| improve this answer | |

HDD can run in a range of 0-60°C, but it is not optimal. They can function with reduced lifespan at either end of these extremes. prolonged exposure under 20°C or above 50°C will cause impairment of the drives useful life. The safest is between 20-45°C. So 38°C is somewhere in this range. 30-32.5°C would be ideal.

One test by national instruments found that a 5C increase could reduce the life by 2 years.

More info here :- https://www.akcp.com/blog/how-temperature-affects-it-data-storage/

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.