Comment 6 for bug 680353

On 24 November 2010 20:02, Andreas Hasenack <email address hidden> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 11/23/2010 09:55 PM, Martin Pool wrote:
>
>> ... or, perhaps better, just use the abstraction across various
>> methods provided by libsensors?  People might also want to graph hdd
>> temperatures?
>
> How does that abstraction work? Can it tell what each "tempN" above is?
> If yes, any particular reason why "sensors" doesn't use it?

But I think 'sensors' does use libsensors. At any rate its package
lm-sensors does depend on libsensors, and it can for example show hdd
temperatures. I haven't looked at the code so I would only be
guessing how it works.

At least on my machine it doesn't know where the tempN probes are.

>>> One of the concerns I have is that the output of sensors is not always
>>> accurate and can be confusing, with names we have no idea what they
>>> represent, like temp1, temp2, etc.
>>
>> I agree, though this does seem to be a problem for acpi too?
>
> Not in our experience, the thermal zones from acpi have been correct so
> far both in values and in which component is being measured.

I seem to recall just often seeing numbered zones in acpi, but you've
probably seen more different machines for this than I have.

>> Another problem, across all methods, is that some sensors seem stuck
>> at unreasonable values like 0 or 127C.
>
> Yes. Unfortunately that coincides with the maximum possible value for an
> 8 bit value, right?

In a way that's fortunate because you can easily conclude it's not a real value.

>> If I was going to programmatically reduce it to  a single value, I
>> would probably take the maximum plausible value at any moment across
>> all sensors.  Or do this per grouping, if we're getting any useful
>> grouping metadata: the hottest probe in the drives is currently: 46C;
>> the hottest probe in the cpu is 48C; etc.
>
> Yeah, but sensors doesn't tell us which temp is cpu, which temp is hard
> drive, etc. Unless the abstraction you mentioned above does.

When I run 'sensors' on my laptop, it does seem to include hdd
temperatures, labelled as such. (Or maybe it's the motherboard sensor
next to the hdd.)

I guess for drives what would be good is to run hddtemp; that could be
a separate bug:

mbp@grace% sudo hddtemp /dev/sd?
/dev/sda: ST31000528AS: 39°C
/dev/sdb: WDC WD10EACS-00ZJB0: 44°C
/dev/sdc: WDC WD10EACS-00D6B0: 47°C
/dev/sdd: WDC WD5000AAKS-22TMA0: 46°C
/dev/sde: WDC WD5000AAKS-22TMA0: 45°C
/dev/sdf: Generic Flash HS-CF: S.M.A.R.T. not available
/dev/sdg: Generic Flash HS-COMBO: S.M.A.R.T. not available

--
Martin