Why does the setting of the kernel parameter kernel.unknown_nmi_panic cause an abnormal restart of a Linux instance? -

This topic describes the causes and solutions for the abnormal restarts after you set the kernel parameter kernel.unknown_nmi_panic.

Problem description

After you set the kernel parameter kernel.unknown_nmi_panic on a Linux Elastic Compute Service (ECS) instance, the instance restarts abnormally. Once the system recovers, the kernel log shows the following call stack information:

[5912557.130943] Uhhuh. NMI received for unknown reason 20 on CPU 0.

[5912557.131115] Do you have a strange power saving mode enabled?

[5912557.131287] Kernel panic - not syncing: NMI: Not continuing

Cause

The issue arises after the kernel parameter kernel.unknown_nmi_panic is set to 1. It leads to a kernel panic when the kernel receives an unknown Non-Maskable Interrupt (NMI), which causes the instance to restart.

The kernel parameter kernel.unknown_nmi_panic controls the behavior of the kernel when it encounters an unknown NMI. The value of kernel.unknown_nmi_panic is set to 0 on most Linux distributions, which means any unknown NMI is ignored.

kernel.unknown_nmi_panic = 1: The kernel will consider a system error and trigger a kernel panic when it encounters an unknown NMI, causing unexpected instance restarts or service interruptions.
kernel.unknown_nmi_panic = 0: The issue is ignored without further action, and the instance continues running in its current state.

Solution

For solution, you can set the kernel parameter kernel.unknown_nmi_panic to 0 to ignore unknown NMIs.

Log on to the ECS instance.
For more information, see Connect to a Linux instance by using a password or key.
Run the following command to check whether any configurations are related to the kernel parameter kernel.unknown_nmi_panic.
```
 grep 'kernel.unknown_nmi_panic' /etc/sysctl.conf /etc/sysctl.d/*.conf
```
- If related configurations exist, delete the redundant configurations and keep only one entry in the /etc/sysctl.conf file. Set it to kernel.unknown_nmi_panic = 0.
- If no related configuration exists, run the following commands to add kernel.unknown_nmi_panic = 0 to the end of the /etc/sysctl.conf file:
```
sudo sh -c "echo >> /etc/sysctl.conf"
sudo sh -c "echo 'kernel.unknown_nmi_panic = 0' >> /etc/sysctl.conf"
```
Run the following command to apply the configuration in the current environment:
```
sudo sysctl -p
```
Note
If you want to modify the kernel parameter without reloading the kernel parameter configuration file, run the command sysctl -w kernel.unknown_nmi_panic=0 to change the value of the kernel parameter kernel.unknown_nmi_panic to 0.