Need help to find out what caused system crash

[Manish Jain]

On 11/12/16 17:30, freebsd-questions-request at freebsd.org wrote:
> Your best bet is to raise a PR with all the detail you can.
> freebsd-questions at ... tends not to be read by all of the kernel
> developers who might be in a position to help you out.
>
> If you can, try and generate a backtrace according to the instructions in
>
> https://www.freebsd.org/doc/en_US.ISO8859-1/books/developers-handbook/kerneldebug-gdb.html


Hi Matthew,

Thanks for replying.

1) As for raising a PR, do you mean I should create an issue at the 
FreeBSD bugzilla ?

2) The URL you have suggested for generating a backtrace says I have to 
use kernel.debug. But 'locate kernel.debug' produces no output on my system.

3) In case I cannot raise an effective PR, I am trying to guess what 
could have been the problem.

This machine is recovering from the effects of a supremely 
malfunctioning graphics card (NVidia, for which the OEM Gigabyte has 
generously declined any support for anything other than Windows).

Replacing that card with a brand new Sapphire Radeon has greatly 
improved things. This system worked well for 2 days straight with the 
new card, and then suddenly crashed once. Just yesterday, I noticed that 
the CPU temperature was mind-boggling. Even with low-to-moderate load, 
my system was generally in the 80 deg Celsius band, with 82 being 
touched every now and then. I did not want that situation to continue, 
so I ordered a new Cooler Master cooler+fan. Installing the new fan a 
few minutes back has brought down CPU temperature crashing down to 46 deg C.

It will be a few days before I can be sure whether the fan switch has 
improved this machine's prospects.

a) Do you think the high CPU temperature could have been responsible for 
the crash ?

b) Is there an application under ports which can display CPU temperature 
? I don't want to be rebooting the system just to have a look into the 
BIOS stats for that.

Thanks again
Manish Jain