You are not logged in.
I'm trying to diagnose some random kernel hangs and to do so I was trying to get a crash kernel setup following the guide here.
Problem is that I can't seem to get a crash kernel to work.
So to test if kexec was having problems on my system I launched the crash kernel normally from kexec with the exact same parameters that I was passing the real crash kernel and everything worked just fine (well not quite. turns out my graphics card was causing some separate issues so I removed it entirely to test. Also it obviously fails to find the dead kernel so the kdump-save service fails).
At first you might be thinking "You probably didn't actually have the crash kernel loaded! Check /sys/kernel/kexec_crash_loaded".
Well actually /sys/kernel/kexec_crash_loaded reports 1 so I do indeed have the crash kernel loaded just fine and no errors are reported loading it manually with kexec.
So I figured something was going wrong early enough in the crash kernel's startup that it never got around to initializing the display before crashing.
I happened to have a FTDI USB to UART converter and a raspberry pi sitting around so I hooked it up with the usb going into the pc and the uart pins hooked up to the pi.
I rebuilt the kdump kernel on my pc to have usb drivers builtin and enabled early kprint over usb then adjusted the crash kernel's command line to add "console=ttyUSB0,115200 console=tty0".
This got me early startup info over usb to my pi.
Unfortunately after setting everything up and crashing the pc with "echo c > /proc/sysrq-trigger" the last thing I see over the uart is something like "Sys-Rq : Triggering a crash".
At this point I'm not sure how to debug this. Any help would be greatly appreciated.
Info:
CPU: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
Motherboard: MSI MS-7850/H97 PC Mate(MS-7850), BIOS V5.9 02/16/2016
Graphics Card (REMOVED): Asus GTX660 TI
RAM: 16GB
Last edited by Fullmetal5 (2018-06-20 03:09:06)
Offline