0

Here is the situation:

I have a Lenovo ThinkCentre M900 with Linux CentOS 7.

I tried to do a kernel load (kexec -l + kexec -e), and it works fine, with one problem.

As soon as I enter the kexec -e command, the Server starts a long alarm beep which sadly doesn't stop.

Am I doing something wrong ? How to reload the kernel without really rebooting ?

Samuel PE
  • 48
  • 5
  • Retry but this time manually unload specific modules before kexecing? (I would start with `modprobe -v -r pcspkr`) – anx Aug 23 '21 at 18:29
  • You probably want to clarify whether CSM (legacy boot) is disabled and whether you could find any other differences pre/post restart of identical kernel (is a PCIe device failing to show up? anything odd about FAN speed control?) – anx Aug 23 '21 at 18:32
  • What would unloading modules do ? Isn't the kexec reloading all modules anyway ? – Samuel PE Aug 24 '21 at 12:32
  • Divide and conquer. If you see one modules unload code is triggering your problem, or if only specific concurrent actions trigger the condition, you do not have to look further into what happens *after* the new kernel instance is in control. – anx Aug 24 '21 at 16:10
  • Ah, I see. But I'm loading the same Kernel. So the modules should be the same. The goal is to bypass the machine's long boot, knowing that some servers take 5 minutes to reach the OS boot. – Samuel PE Aug 26 '21 at 07:15

0 Answers0