Configuring kdump in RHEL

kdump may be very useful for you if you're experiencing kernel crashes and want to analyze what is really happening there (or you may pass this part to support).
First install kdump package:
# yum -y install kexec-tools
Then modify /etc/grub.conf and add "crashkernel=128M@16M" after "root=LABEL=/" for current kernel.

It's necessary to reboot the box to make the new settings active:
# /sbin/init 6

After the reboot make sure that kdump service is enabled and running:
# chkconfig kdump on
# service kdump start

You should also test if your setup is fine. Open console and enter the following command to trigger the dump:
echo "c" > /proc/sysrq-trigger