Troubleshooting kernel issues

Note: The hypervisor used by EC2 prevents kdump from being engaged. Forcing a kernel crash will simply prevent you from logging back in for 3-4 miniutes. No crash will be present in /var/crash.

---

Start and enable the kdump service:

systemctl start kdump && systemctl enable kdump

You'll notice kdump doesn't start. Install necessary packages:

yum --enablerepo=rhel-7-server-rhui-debug-rpms install -y kexec-tools crash kernel-debug kernel-debuginfo-$(uname -r)

Modify /etc/default/grub to set crash memory to 128M.

Install the new Grub config:

grub2-mkconfig -o /boot/grub2/grub.cfg

Install SystemTap:

yum install -y systemtap kernel-devel-$(uname -r)

Run the example script: /root/example.stp

stap /root/example.stp

Compile the script as a module:

stap -p 4 -v -m stap_example example.stp

Create a systemtap directory for the module:

mkdir /lib/modules/$(uname -r)/systemtap

Move the module into the new directory:

mv stap_example.ko /lib/modules/`uname -r`/systemtap/

Run the module:

staprun stap_example

Add cloud_user to the stapusr group:

usermod -aG stapusr cloud_user

Verify cloud_user can run the module:

runuser -l cloud_user -c 'staprun stap_example'