Burning drives
#284 Henry, Monday, 21 September 2015 11:57 AM (Category: Hardware)
(Tags: smartd harddisk)

I use smartd to monitor the health of my hard disks. I've been seeing notices like this on a regular basis:

Sep 21 10:57:10 cerebus smartd[857]: Device: /dev/sdb [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 115 to 114

where the temperature of the hard disk goes up one degree, then later it goes down one degree. It's happening on three different servers.

But something struck me - it's called Temperature_Celsius and it's running at 114 degrees?

I opened my desktop up while it was running, and touched the drive. My hand did not crisp and burn. There is no way it is at 114 degrees Celsius. I don't understand. Is it called Temperature_Celsius but it's really in Fahrenheit?

I Googled for it and got this page which might explain it.

And I should have looked at /etc/smartd.conf too. The first section that sets DEVICESCAN provides a number of options that handle the temperature:

# Alternative setting to ignore temperature and power-on hours reports
# in syslog.
#DEVICESCAN -I 194 -I 231 -I 9

# Alternative setting to report more useful raw temperature in syslog.
#DEVICESCAN -R 194 -R 231 -I 9

# Alternative setting to report raw temperature changes >= 5 Celsius
# and min/max temperatures.
DEVICESCAN -I 194 -I 231 -I 9 -W 5

I went with the last one, only wanting to know about big jumps in temperature. So now at startup, I get an initial valuation of the raw temperature.

Sep 21 13:21:39 cerebus smartd[7317]: Device: /dev/sdb [SAT], initial Temperature is 36 Celsius (Min/Max ??/36) 

And 36 Celsius is what my fingers tell me is about right, and I'm okay with that.