Overview
When a task or kernel thread is using and not releasing a CPU for a period of time, it will generate a crash dump on a soft lockup that you will need to provide to the support agent for further troubleshooting.
This article explains how to configure the server to generate the core dump files required to troubleshoot the issue.
Solution
To configure the server to generate the core dump files, ssh into the server as the root user, and follow the steps below:
- Get the current configuration:
cat /proc/sys/kernel/softlockup_panic
- Enable the kernel panic (when a soft lockup is detected) with the command below:
echo 1 > /proc/sys/kernel/softlockup_panic
- Verify the kernel panic has been enabled by running the command below:
cat /proc/sys/kernel/softlockup_panic
- Ensure the setting survives a reboot by adding the below line to the
/etc/sysctl.conffile:
kernel.softlockup_panic=1