During Upgrade of Oracle Big Data Appliance, Node Cannot be Booted into the Latest Kernel Either into OS or USB Mode

Asset ID:	1-72-1980574.1
Update Date:	2018-04-07
Keywords:

Solution Type Problem Resolution Sure

Solution 1980574.1 : During Upgrade of Oracle Big Data Appliance, Node Cannot be Booted into the Latest Kernel Either into OS or USB Mode

Applies to:

Big Data Appliance Hardware - Version All Versions and later
Linux x86-64

Symptoms

Upgrading to BDA Mammoth 4.1 Release. After the upgrade node01 seems to hang on boot.

In ILOM remote console noticing below errors on boot phase ....

find (hd3,0)/THIS_IS_AN_OS_DIS

Error 15: File not found

find (hd3,0)/THIS_IS_AN_OS_DISK

Error 15: File not found

Also there doesn't seem to be USB Rescue boot options in the grub to be able to boot into rescue mode.

Only able to perform OS boot into an old kernel by following Doc ID 1983550.1 Instructions to Manually Enter Boot Options in Grub Shell during Reboot on Oracle Big Data Appliance.

Cause

After booting into old kernel execute 'rpm -qa | grep kernel-uek' to check the kernels installed.

# rpm -qa | grep kernel-uek

..........
kernel-uek-debug-2.6.39-400.215.1.el5uek
kernel-uek-2.6.39-400.214.3.el5uek
kernel-uek-firmware-2.6.39-400.215.1.el5uek
kernel-uek-doc-2.6.39-400.215.1.el5uek
.......
kernel-uek-2.6.39-400.215.1.el5uek
kernel-uek-devel-2.6.39-400.215.1.el5uek
kernel-uek-devel-2.6.39-400.212.1.el5uek
kernel-uek-debug-2.6.39-400.214.3.el5uek
kernel-uek-headers-2.6.39-400.215.1.el5uek

Though rpm -qa command states the latest 215/214 Kernel rpms are installed as part of BDA 4.1/4.0 upgrade found that but vmlinuz-2.6.39-400.215 and also initrd-2.6.39-400.215/214 files are missing from /boot directory

Also executing 'rpm -V' command returns file missing error:

# rpm -V kernel-uek-2.6.39-400.215.1.el5uek

missing /boot/vmlinuz-2.6.39-400.215.1.el5uek

Solution

1) Copy the missing /boot/vmlinuz-<KernelVersion> files from working node to failing node.

2) Copy the missing /boot/initrd-<KernelVersion> files from working node to failing node.

3) Check rpm -V kernel-uek-<KernelVersion> to ensure no errors.

4) Now reboot the node so the node can be booted into latest kernel.

Attachments

This solution has no attachment