logging heat and power etc.

[Cr0n_J0b]

My system has shut down a few times with no warning...always when I'm away. It just locks up. I think it might be heat related but I can't figure out how to get my lm_sensors to keep a log of the heat and power measurements. Any one have an idea here? I think this is pretty basic, but i searched google and the lm_sensors FAQ and couldn't find it.

thanks

[peters]

Try running 'sensors' by itself and see what it gives you. I wrote some scripts on my little web server that actually charts out and graphs harddisk temperature and the MB temperature using a bit of this data. I think I'm pulling the HD temp using 'hdparm' . You can do 'crontab -e' to create a cronjob (UNIX 101). Hope you are familiar with Vi. There are probably other tools to do it. You could probably, every 5 minutes, append the temps you want to to a file somewhere along with the date and time.. Pretty easy stuff. And fun. Btw, if your system spontaneously shuts down I recommend running "data=journal" mode on your ext3's. Look at 'tune2fs' and '/etc/fstab' to change your settings to enable that. What shows up in your logs? Use 'less /var/log/messages' and see if there are any warnings that occur before shutdown. That is pretty weird.

[Cr0n_J0b]

I'll take the scripting under advisement. I was hoping to find something off the shelf that could just do this. I think that sensord should do this, but it's not included with lm_sensors anymore. I was going to load ip up, but there are a bunch of dependend files that I would have to load first.

I checked messages and I'm seeing lots of firewall hits, but that's all up to the lockup at 5:00am. There's some hard drive errors that I'm going to need to look into on one of my raid sets, but I don't think that's it...maybe...

I moved around some of the cards in my case to give a bit more breathing room for my video card. It's a high performance card and didn't really have room to vent...I'm hoping that was the issue.

The CPU is also running hot and I don't know why. It's at about 51C, but the case is at 30C and it's got a huge fan and HS on it. I tried resetting it and taking off some of the thermal paste, thinking that I might have had too much, but it's still still hotter than I would like...it should be in range for operation however.

one last question:

there is a command that I've used in the past that will echo everything going to the message file to the terminal...but i forget the syntax. Any thoughts?

[duncan]

tail -f /var/log/messages

[peters]

I like to do "less -Sn /var/log/messages", then once I'm in I type "F" ... Same thing as tail, but I can CTRL-C and then look at the past.. And reconnect. Etc.

[simonn]

I just built a quiet system and to test the temperature I wrote a script to write the date/time, M/B and CPU temperatures from sensors to a log and called it every minute via cron:

the script /root/testtemp:

date >> /root/temp.log
sensors | grep Temp >> /root/temp.log

/etc/crontab

SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
HOME=/

# run-parts
01 * * * * root run-parts /etc/cron.hourly
02 4 * * * root run-parts /etc/cron.daily
22 4 * * 0 root run-parts /etc/cron.weekly
42 4 1 * * root run-parts /etc/cron.monthly
*  * * * * root /root/testtemp

Once this is setup, run cpuburn-in for a few hours.

What CPU is it BTW? Some do run that hot (or hotter e.g. AMD XP).

I suspect that thermal paste is the cause, at least it was for me with my old XP 2600 (now have Sempron 64 3000 which is a lot cooler). Remove all the current paste with ArctiClean and reapply some arctic silver or something. You really do not need much.

Also, what kernal are you running? 2.6.15-1.1830_FC4 Caused my PC to do a similar thing when I had my external HDD plugged in. 2.6.15-1.1831_FC4 fixed it.