Hddetemp vs hddtemp

[Markus Robert Kessler]

Hi everyone,

maybe this info is helpful for someone:

With "hddtemp" there's a tool out, capable of monitoring harddisks'
temperature. This may help to prevent them from heat death.

Unfortunately, hddtemp (every version I know of) can only access internal
drives. External ones can not be read. Don't know why.

So, I developed a little workaround and called it hddetemp (e for
external, also). It is now also tested under Raspbian OS.

Here you see the difference:

[542 root@rpi-inst /]# hddtemp /dev/sda
/dev/sda: Generic External: S.M.A.R.T. not available

[543 root@rpi-inst /]# hddetemp /dev/sda
/dev/sda: Hitachi HTS545016B9A300: 25 (Min/Max 16/43)°C

If you find this helpful, you can get it from here:

https://www.dipl-ing-kessler.de/developer/freigabe/hddetemp

Best regards,

Markus


--
Please reply to group only.
For private email please use http://www.dipl-ing-kessler.de/email.htm

[Marco Moock]

Am Samstag, 17. September 2022, um 17:20:51 Uhr schrieb Markus Robert
Kessler:

> If you find this helpful, you can get it from here:
>
> https://www.dipl-ing-kessler.de/developer/freigabe/hddetemp

Great, but wouldn't it be the best solution to implement your
extensions into hddtemp?
Did you contact the original developer about this?

[Jan Panteltje]

On a sunny day (Sat, 17 Sep 2022 17:20:51 -0000 (UTC)) it happened Markus
Robert Kessler <no_r...@dipl-ing-kessler.de> wrote in
<tg4vlj$77k3$1...@dont-email.me>:

>Hi everyone,
>
>maybe this info is helpful for someone:
>
>With "hddtemp" there's a tool out, capable of monitoring harddisks'
>temperature. This may help to prevent them from heat death.
>
>Unfortunately, hddtemp (every version I know of) can only access internal
>drives. External ones can not be read. Don't know why.
>
>So, I developed a little workaround and called it hddetemp (e for
>external, also). It is now also tested under Raspbian OS.
>
>Here you see the difference:
>
>[542 root@rpi-inst /]# hddtemp /dev/sda
>/dev/sda: Generic External: S.M.A.R.T. not available
>
>[543 root@rpi-inst /]# hddetemp /dev/sda

>/dev/sda: Hitachi HTS545016B9A300: 25 (Min/Max 16/43)°C

>
>If you find this helpful, you can get it from here:
>
> https://www.dipl-ing-kessler.de/developer/freigabe/hddetemp
>
>Best regards,
>
>Markus

There is the main HD info program
hdparm
man hdparm

# hdparm -H /dev/sdb
/dev/sdb:
drive temperature (celsius) is: under -20
drive temperature in range: yes

That is with my external 3 TB Toshiba drive
Not sure 'under -20' makes any sense.

I tried your code, installed 'smartctl':

# hddetemp /dev/sdb
/dev/sdb: TOSHIBA MQ04UBB400: °C

So.. may work for some drives.....

[Paul]

smartctl of smartmontools, knows the temperature.
It looks at location 190 and 194, depending on drive type.
Some devices will require smart passthru, to gain access.

suso smartctl --all /dev/sda | grep -i temperature # basic idea, remove the grep and review by hand

And gnome-disks seems to as well, got both drives (local sata, USB SSD),
It would be interesting to see what cluster of utilities gnome-disks uses.
This is my go-to tool for daily usage.

inxi -F seems a bit lame in the temperature department, getting one of two drives.

The hardware has all sorts of quirks, so this is a topic
for individuals with the time to spend on it.

Paul

[Markus Robert Kessler]

Hi Marco,

implementing this code into hddtemp source would be nice, but looking at

http://download.savannah.nongnu.org/releases/hddtemp/

(official homepage) it seems that hddtemp is no longer maintained.
Updating the drives database is not enough to get ahead here.

Thanks, best regards,

[Markus Robert Kessler]

Hi Jan,

could you please paste smartctl's output for your drive,
section "Vendor Specific SMART Attributes with Thresholds:",
like

ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE
UPDATED WHEN_FAILED RAW_VALUE
[...]
190 Airflow_Temperature_Cel 0x0032 068 049 000 Old_age
Always - 32
[...]

Thanks!

[Daniel65]

Markus Robert Kessler wrote on 18/9/22 6:56 am:

> On Sat, 17 Sep 2022 19:58:10 +0200 Marco Moock wrote:
>
>> Am Samstag, 17. September 2022, um 17:20:51 Uhr schrieb Markus Robert
>> Kessler:
>>
>>> If you find this helpful, you can get it from here:
>>>
>>> https://www.dipl-ing-kessler.de/developer/freigabe/hddetemp
>>
>> Great, but wouldn't it be the best solution to implement your extensions
>> into hddtemp?
>> Did you contact the original developer about this?
>
> Hi Marco,
>
> implementing this code into hddtemp source would be nice, but looking at
>
> http://download.savannah.nongnu.org/releases/hddtemp/
>
> (official homepage) it seems that hddtemp is no longer maintained.
> Updating the drives database is not enough to get ahead here.
>
> Thanks, best regards,
>
> Markus
>
>

Is Hddtemp just a spinning drive-type tool or is it applicable for Solid
State Drives as well??
--
Daniel

[Marco Moock]

Am Sonntag, 18. September 2022, um 23:09:24 Uhr schrieb Daniel65:

> Is Hddtemp just a spinning drive-type tool or is it applicable for
> Solid State Drives as well??

Why should that matter?
There is a temperature sensor that need to be read out, regardless of
the technology the disk uses.
But according to the other post hddtemp is very old, so it is from a
time where SSD weren't common.

[Henry Crun]

What works for me (on SSD):

nvme smart-log /dev/nvme0 | grep temperature

from "man nvme" ;

<quote>
NVME(1) NVMe Manual NVME(1)

NAME
nvme - the NVMe storage command line interface utility (nvme-cli)

SYNOPSIS
built-in plugin:

nvme <command> <device> [<args>]
</quote>




--
No Micro$oft products were used in the URLs above, or in preparing this message.
Recommended reading: http://www.catb.org/~esr/faqs/smart-questions.html#before
and: http://alpha.mike-r.com/jargon/T/top-post.html

[Paul]

For reasons that escape pretty well everyone, the SMART table on
legacy HDD and on the new SSD, don't match. You would think there
would be value, having entries with the same name, at their
traditional address. I would guess that is too easy.

Some environment information is pinned out in ACPI tables, such as
CoreTemp from your CPU. One of the improvements in reading the
hardware monitor interface on the SuperIO, was for the BIOS to
wrap some of the information, making it easier for userland applications
to "consume" the information. For example, the scaling resistors on
the voltage measurement section, if you use the ACPI table, the
voltages are already scaled for you. And that saves an unbelievable
amount of labor. In the past, humans used to submit empirically
derived scale resistor info, to the maintainer of MBM5, per motherboard
design. And that sucked as a method. Having the correct values
computed by the BIOS designer, and passed via ACPI table, made
so much more sense.

But not everything in life is that easy. Thus, we're still fiddling
with SMART tables in the year 2022.

I assume NVMe temperature is the same as SATA SSD, but that would
be a rash assumption unless verified. Even an eMMC chip, could have
its temperature information, delivered in some other way. And USB
flash sticks, I've never seen a measured value printed on a screen,
for those.

Paul

[Marco Moock]

Am Sat, 17 Sep 2022 17:20:51 -0000 (UTC)
schrieb Markus Robert Kessler <no_r...@dipl-ing-kessler.de>:

> If you find this helpful, you can get it from here:

Are there any plans to provide deb packages and/or include this
software in the official repositories of Debian/Ubuntu?

Then more people will likely use it.

[William Unruh]

Tried it on a system with and external drive inside a usb converter, and
the systemctl command simply hangs. ^C cannot free it, but
sudo killall hddetemp
did kill it.

[Markus Robert Kessler]

On Wed, 21 Sep 2022 13:53:07 +0000 William Unruh wrote:

If you look into the source, you see that hddetemp has nothing to do with
systemctl. It relies on smartctl. If this hangs, then something is wrong
with your hardware.