I have Oracle Enterprise Linux 6.6 installed and up-to-date with the latest Unbreakable Enterprise Kernel installed. The default kernel is set to 3.8.13-68.el6uek.x86_64, with a crash kernel set to 2.6.32-504.12.2.el6.x86_64. The machine will crash to the 2.6 kernel every time. I don't know why, or how to debug this.

Any tips on how to debug this?

I've figured it out. The situation is highly specific, so here is more information and my solution:

Environment:
Oracle Linux 6.6 virtual machine w/ UEK 3.8
XenServer 6.5 hypervisor (previously upgraded from 6.2)

Explanation:
XenServer 6.2 does not support UEK and attempting to boot a VM with this kernel can cause the entire host to crash. Therefor the UEK kernel is blacklisted at the Dom0 level, to prevent it from booting in any DomU. XenServer 6.5 does support UEK, however if the installation has been upgraded from 6.2, the blacklist rules persist and must be manually removed.

Solution:
From the host machine (Dom0):
# rm /etc/pygrub/rules.d/oracle-5.x
# rm /etc/pygrub/rules.d/oracle-5.6
From the VM:
# reboot now
$ uname -r
3.8.13-68.el6uek.x86_64