RHEL: Crash Dump Configuration using NMI in Red Hat Enterprise Linux

Configure crash dump on Red Hat Enterprise Linux. 

Listed below are the steps to configure crash dump on Red Hat Enterprise Linux and NMI functionality in iDRAC:

• Ensure that the CMC, iDRAC, BIOS firmware is up to date.
• Configure Kdump on Red Hat Linux using the following article - How to troubleshoot kernel crashes, hangs, or reboots with kdump on Red Hat Enterprise Linux - Red Hat Customer Portal   
• Allocate sufficient space for the dump so that the entire memory contents are captured in the dump
• Enter the following command to determine whether the system is set to call the panic routine when it receives an unknown NMI:

# sysctl kernel.unknown_nmi_panic 

• If the system is not configured to call the panic routine when it receives and unknown NMI, this command produces the following output:

kernel.unknown_nmi_panic = 0 

• When using the default kernel setting, the Linux kernel logs the occurrence of an unknown NMI in /var/log/messages, but the system does not call the panic routine.
• To enable the kernel to call the panic routine when it receives an unknown NMI, edit the /etc/sysctl.conf file and ensure that the following line is in the file:

kernel.unknown_nmi_panic = 1 

• Enter the following command to reload the setting. You should see the report of the new value as part of the reload output: 

# sysctl -p

net.ipv4.icmp_echo_ignore_broadcasts = 1

net.ipv4.conf.all.rp_filter = 1

kernel.unknown_nmi_panic = 1 

• Once the server has stopped responding, login to iDRAC - Power management - NMI - Apply. The system starts dumping the memory, wait for 30-45 minutes for the dump to complete. When completed the server reboots.

Red Hat Support can perform a thorough dump analysis.