Hi Guys,

Couple of easy questions for you all, we are conducting a basic sort of stress test on SATA drives in our prototype, and to monitor the temperature changes of these drives, smartd seemed a reasonable solution. But there are a few kinks which need to be removed.

1. In smartd.conf, I have the following line (for each drive)
/dev/sdk -d sat -I 194 -W 1,30,50

the goal is to report every temperature change, that is why I have 1 after -W, I don't really need the second one, and I am not sure what is the use of the third one. Is there a better way to do it ??

2. I would like to log the output of smartd, but it seems that output of smartd is being logged in /var/log/messages, is there way to specify output file for smartd?? If not, is there a work around??

3. In the following output, why sometimes the smartd was unable to read temperature??

Thanks


Nov 15 12:03:35 patrick smartd[17487]: Device: /dev/sdk, failed to read Temperature
Nov 15 12:03:35 patrick smartd[17487]: Device: /dev/sdg, failed to read Temperature
Nov 15 12:04:31 patrick smartd[17524]: Device: /dev/sdk, Temperature changed +1 Celsius to 30 Celsius (Min/Max 30/30!)
Nov 15 12:04:31 patrick smartd[17524]: Device: /dev/sdk, Temperature 30 Celsius reached limit of 30 Celsius (Min/Max 30/30!)
Nov 15 12:04:31 patrick smartd[17524]: Device: /dev/sdg, Temperature changed +1 Celsius to 32 Celsius (Min/Max 32/32!)
Nov 15 12:04:31 patrick smartd[17524]: Device: /dev/sdg, Temperature 32 Celsius reached limit of 30 Celsius (Min/Max 32/32!)
Nov 15 12:13:02 patrick smartd[16708]: Device: /dev/sdk, failed to read Temperature
Nov 15 12:13:03 patrick smartd[16708]: Device: /dev/sdk, failed to read Temperature
Nov 15 12:33:36 patrick smartd[17487]: Device: /dev/sdk, failed to read Temperature
Nov 15 12:33:36 patrick smartd[17487]: Device: /dev/sdg, failed to read Temperature
Nov 15 12:34:31 patrick smartd[17524]: Device: /dev/sdk, Temperature changed -1 Celsius to 29 Celsius (Min/Max 29!/30)
Nov 15 12:34:31 patrick smartd[17524]: Device: /dev/sdg, Temperature 32 Celsius reached limit of 30 Celsius (Min/Max 32/32)
Nov 15 12:43:02 patrick smartd[16708]: Device: /dev/sdk, failed to read Temperature
Nov 15 12:43:03 patrick smartd[16708]: Device: /dev/sdk, failed to read Temperature


--
Saqib bin Sohail
University of Colorado at Boulder
(303) 786 0636
http://ucsu.colorado.edu/~sohail/index.html