How to log Linux kernel loading?

Question

linux

There is a system on Centos 6.5, a kernel update was recently released from the official repository:
2.6.32-431.17.1
Now the kernel version is a little older, about March, again from official sources.
One day, when I ran `yum update`, yum prompted me to update the kernel. Right after the upgrade, I took a look at the grub config to make sure the new kernel would boot.
I execute the reboot command, but the server does not boot.
The server is remote and there is no KVM access. There is only Rescue and NetBoot. In the /var/log/boot.log and dmesg files, I only see the logging of the last successful boot of the system.
How to correctly log kernel startup on Centos 6.5 to find the reason why the kernel does not want to start?

Answer 1

For good, there is no way without a remote KVM, since a failure can occur at the very beginning of the boot, for example, in the kernel driver or in the initrd.img scripts, or the boot-record simply fails!
The most correct way is to install a BMC module with remote management support on the server, which allows you to watch not only the console, but also remotely connect a CD to it, redirect the serial-console, turn on / off the server remotely, and much more.
Almost all server boards support this functionality either directly (through the BMC controller) or by installing an additional patch scarf.