Confused about SMART info and HDD temps

[akovano]

I know I can check hdd temp by running

Code:

smartctl -a /dev/da[x]

but I am unclear if this is pulling the current temperature value, or if this is the disk temperature when a SMART test was last run?

I wanted to keep an eye on my disk temperatures as the ventilation around my rack is not ideal.
I have SMART test (short) scheduled to run once a week, so I wasn't clear if smartctl was looking at the result of the previous test or pulling the current temperature value.

 

[jgreco]

It's whatever the drive manufacturer decided to do. Generally it represents some sort of "current" temperature, but who knows how often it is updated.

 

[Alecmascot]

check it with a -a then run the short test "-t short" will take 1 or 2 minutes, then check it again.

 

[akovano]

Thanks, I tried this for one of the drives and got the same temp result.

So when "smartctl" is run, is this actually querying the drive directly for a result (which may or may not be current depending on the manufacturer)?

Or is this command getting the result from a previous test result (stored in TrueNAS)?

Basically I'm trying to figure out if I need to run a SMART test every time I want to check the temperatures, or if I can have a script that greps the result from smartctl for each drive every so often and that will give a somewhat current result.

 

[jgreco]

o when "smartctl" is run, is this actually querying the drive directly for a result

smartctl is interactively chatting with the drives. What that conversation consists of may be gibberish on the HDD's side, and a bunch of very clever people have put a lot of effort into smartctl over the years to try to decipher it and make it usable and at least somewhat useful.

 

[Ericloewe]

Basically I'm trying to figure out if I need to run a SMART test every time I want to check the temperatures

That is definitely not the expectation.

or if I can have a script that greps the result from smartctl for each drive every so often and that will give a somewhat current result.

Yeah, that's more in line with disks seem to do. I say "seem" because they can be rather opaque (to say nothing of outright lies).

 

[akovano]

OK understood thanks.
FWIW, I just ran a test on this. Opened up the case and blew a steady stream of air across the drives for 5-10 minutes.
Without running a SMART test, I kept checking the results of smartctl and after a couple minutes a saw a slight drop in temp. So it does pull somewhat current values. Good to know.

 

[Ericloewe]

The -x option should also include a time-series of temperature values recorded by the disk, which is occasionally useful to understand patterns.