My laptop crashed - maybe heat? - How can I get logs or warned?

Is there a way to get a warning when the computer will crash
(perhaps due to heat) or at least read a log file that tells me why?

In the past month, at least twice, maybe 3 times of constant running,
my laptop just went blank in front of me, and shut down.

I don't know why. I only know to do a "dmesg" but when I reboot and
then run "dmesg", I only get messages that I don't fully understand
or which don't seem related to the crash, for example:

$ dmesg
[16554.569776] wlan0: authenticated
[16554.570992] wlan0: associate with 88:24:01:44:a1:23 (try 1/3)
[16554.579745] wlan0: associate with 88:24:01:44:a1:23 (try 2/3)
[16554.589676] wlan0: associate with 88:24:01:44:a1:23 (try 3/3)
[16554.598817] wlan0: RX AssocResp from 88:24:01:44:a1:23 (capab=0x11 status=0 aid=1)
[16554.603672] wlan0: associated
[16554.603713] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[16554.632817] wlan0: deauthenticating from 88:24:01:44:a1:23 by local choice (reason=2)
[16554.640791] cfg80211: Calling CRDA to update world regulatory domain
[16554.640873] wlan0: authenticate with 88:24:01:44:a1:23
[16554.645541] wlan0: send auth to 88:24:01:44:a1:23 (try 1/3)
[16554.646092] cfg80211: World regulatory domain updated:
[16554.646099] cfg80211: (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
[16554.646107] cfg80211: (2402000 KHz - 2472000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[16554.646113] cfg80211: (2457000 KHz - 2482000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[16554.646118] cfg80211: (2474000 KHz - 2494000 KHz @ 20000 KHz), (300 mBi, 2000 mBm)
[16554.646123] cfg80211: (5170000 KHz - 5250000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[16554.646129] cfg80211: (5735000 KHz - 5835000 KHz @ 40000 KHz), (300 mBi, 2000 mBm)
[16554.655366] wlan0: send auth to 88:24:01:44:a1:23 (try 2/3)
[16554.657113] wlan0: authenticated
[16554.658945] wlan0: associate with 88:24:01:44:a1:23 (try 1/3)
[16554.667039] wlan0: associate with 88:24:01:44:a1:23 (try 2/3)
[16554.676943] wlan0: associate with 88:24:01:44:a1:23 (try 3/3)
[16554.677940] wlan0: RX AssocResp from 88:24:01:44:a1:23 (capab=0x11 status=0 aid=1)
[16554.682145] wlan0: associated

Is there a way to get a warning when the computer will crash
(perhaps due to heat) or at least read a log file that tells me why?

 

yeah, likely there is an addon for your GUI for monitoring the heat of
your laptop or desktop. Likely need the lm-sensors package as part of
that. And it wouldn't hurt to clean out the fans and make sure the
machine can breath well.

I've not toyed with this myself. If there is a command line for it, then
a script could be made to tap the sensor once and a while and record the
data to a file so you can view it after a crash and see what had been
going on.

 

Install rsyslog, then you can check /var/log/syslog. Before I got a
really good cpu cooler, I had occasional shutdowns caused by heat.
I now have a corsair h55 cooler. While the sensors command shows the
cpu staying cool, their can sometimes be a lot of heat coming from
the radiator. Good in winter, not so good in summer. Note that the
cooler requires a fairly wide, tall case. I've been using cooler
master haf 912 cases for the last year or two.

For a laptop, cleaning the system is paramount. Take it apart if
needed, to clean properly.

Regards, Dave Hodgins

 

In /var/log I curently have the following *.log files (skipping
numerical duplicates):
- alternatives.log
- apport.log
- auth.log
- boot.log
- bootstrap.log
- dmesg
- dpkg.log
- fontconfig.log
- gpu-manager.log
- jockey.log
- kern.log
- lastlog
- mail.log
- pm-powersave.log
- prime-offload.log
- prime-supported.log
- syslog
- Xorg.0.log

I just installed rsyslog using:
$ sudo apt-get install rsyslog

It didn't immediately put a log file in there, and,
running the command didn't work. It doesn't seem to
have a manpage.

Googling, I find this page, where the "r" apparently
stands for "rocket":
http://www.rsyslog.com/doc/master/index.html

 

I'm surprised the temperature fluctuates as much as it does given
the laptop is in the same spot indoors all the time.

Here is the laptop temperature after sitting idly all night:
$ sensors
acpitz-virtual-0
Adapter: Virtual device
temp1: +59.0°C (crit = +97.0°C)

thinkpad-isa-0000
Adapter: ISA adapter
fan1: 3115 RPM
temp1: +59.0°C
temp2: +0.0°C
temp3: +0.0°C
temp4: +0.0°C
temp5: +0.0°C
temp6: +0.0°C
temp7: +0.0°C
temp8: +0.0°C

coretemp-isa-0000
Adapter: ISA adapter
Core 0: +58.0°C (high = +84.0°C, crit = +100.0°C)
Core 1: +58.0°C (high = +84.0°C, crit = +100.0°C)
Core 2: +60.0°C (high = +84.0°C, crit = +100.0°C)
Core 3: +59.0°C (high = +84.0°C, crit = +100.0°C)

 

That page, to me, was almost useless, as it didn't give me
a tutorial of how to use rsyslog.

Apparently people have had the same problem with rsyslog:
http://askubuntu.com/questions/562505/how-to-use-rsyslog

That article implies the "r" is for "remote" (not "rocket").

 

It is the cpu that heats it up. The cpu uses a lot of energy when
working, esp older cpus. But 90 deg is extreme, and suggest that there
is some cooling problem in your laptop.
Heat sink? fan? cooling slots plugged?

 

I keep it elevated in the back on a quarter-inch block of wood, so,
it's even more open than it would be.

But, I think you guys are right that heat is what is killing the laptop.

What I need is a command that *warns* me when the temperature gets close
so that I can prepare to shut down running programs to save data.

This pretty much tells me the temperature (those are single quotes in the awk)
# alias temp='sensors | grep "Core 0:" |awk '{print $3}''

What I need to look up is how to do the "if" so that it spits out a
warning when the temperature gets close to a set point (such as 84 degrees).

 

What you need is to fix the cooling problems. Once your laptop starts to
crash duw to heat, it will get worse and worse, as the atoms in the
silicon sart to diffuse out of the areas where they are supposed to be,
creating shorts, etc.

sensord will tell your system.
man sensord

 

$ sensord
The program 'sensord' is currently not installed.

$ sudo apt-get install sensord
Processing triggers for man-db (2.6.7.1-1ubuntu1) ...
Processing triggers for ureadahead (0.100.0-16) ...
Setting up libdbi1:amd64 (0.9.0-1) ...
Setting up librrd4 (1.4.7-2ubuntu5) ...
Setting up sensord (1:3.3.4-2ubuntu1) ...
* Starting sensor daemon sensord [ OK ]
Processing triggers for libc-bin (2.19-0ubuntu6.6) ...
Processing triggers for ureadahead (0.100.0-16) ...

$ man sensord
Sensord is a daemon that can be used to periodically log
sensor readings from hardware health-monitoring chips to
syslog(3) or a round-robin database (RRD) and to alert when
a sensor alarm is signalled; for example, if a fan fails,
a temperature limit is exceeded, etc.

Unfortunately, none of the examples in the manpage pertain
to temperature alarms, so, I'll google further (is it just
me or are manpages almost useless for actually getting something
done that you want to get done for others also?).

 

I suspect "sensord" is the wrong command for the very common
need of setting a temperature alarm, because I googled for how
to set a temperature alarm, and *none* of the references even
listed 'sensord', let alone gave an example of how to set up
a sensord temperature alarm.

https://www.kubuntuforums.net/showthread.php?42895-CPU-temperature-monitoring-(intel-i7)
http://ubuntuforums.org/showthread.php?t=2036321
https://help.ubuntu.com/community/SensorInstallHowto
http://askubuntu.com/questions/tagged/temperature
http://ubuntuforums.org/showthread.php?t=1494180
etc.

Not one of the references above used sensord, so, I'm pretty sure
sensord is not the right solution (even if its manpages were useful).

This seems to be the right solution, but I have to try it first:
http://fosswire.com/post/2007/09/set-up-a-system-tray-temperature-monitor-with-ksensors/
$ sudo apt-get install lm-sensors
$ sudo sensors-detect
$ service kmod start
$ sudo apt-get install ksensors

DRAT!
$ sudo apt-get install ksensors
Reading package lists... Done
Building dependency tree
Reading state information... Done
Package ksensors is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Package 'ksensors' has no installation candidate