Kernel is Not Upgraded After Mammoth Step 2 Reboot When a Kernel with a Lower Version but More Recent Build Time than the Kernel Supplied with the Upgrade is Present
(Doc ID 2308068.1)
Last updated on JANUARY 24, 2020
Applies to:Big Data Appliance Integrated Software - Version 4.4.0 and later
After the Step 2 reboot of a Mammoth upgrade the kernel is not upgraded to the kernel provided with the new Mammoth version.
1. Generally all checks look ok:
a) /boot has sufficient space.
b) /etc/yum.conf looks ok and /etc/yum.repos.d only contains bda.conf.
c) The post_upgrade_script_*.out looks fine.
d) All the /etc/grub, /usbdisk/boot/grub/grub.conf links are correct.
e) "rpm -qa | grep uek" shows the new kernel is installed.
f) All nodes rebooted successful and have a /root/BDA_REBOOT_SUCCEEDED file.
2. However the /etc/grub.conf (which is a link to /boot/grub/grub.conf) does not have the upgraded kernel at the top of the file.
At the top of /etc/grub.conf is a kernel which is not the kernel provided with the current Mammoth upgrade. The kernel provided with the Mammoth upgrade is lower on the list, which is the reason that hosts are not booting into the new kernel.
3. Further examination finds that the kernel at the top of the grub.conf file is a kernel with a lower version but more recent build time than the kernel supplied with the Mammoth upgrade.
4. Consider the example of a Mammoth 4.4. to 4.7 upgrade:
On 4.4 the kernel is:
On 4.7 the kernel is:
But in the case here the grub.conf file has a UEK2 kernel: 2.6.39-400.286.2.el6uek.x86_64 at the top of the file. That kernel is followed by the Mammoth V4.4 default kernel which is followed by the Mammoth upgrade UEK 4 kernel.
Note: 2.6.39-400.286.2.el6uek.x86_64 is not a kernel supplied by Mammoth. It is a kernel that the cluster has been upgraded to.
Checking the build times shows that the kernel 2.6.39-400.286.2.el6uek.x86_64 has a more recent build time than 4.1.12-70.el6uek.x86_64.
If the kernel is not upgraded after Step 2 reboot and the symptoms match those above run the above command to confirm that there is a lower versioned/higher build time kernel on the cluster than that which the Mammoth upgrade is upgrading to.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!
In this Document