Tuesday, December 29, 2015

Kernel panic in linux

Below are the steps to recover the kernal panic issue in Oracle Exadata machines

Take the diagnostics ISO from the working server and map it in faulty server then boot that server using that ISO and follow the below steps

mkdir /mnt/LVDbSys1
mkdir /mnt/LVDbSys2

mount /dev/mapper/VGExaDb-LVDbSys1 /mnt/LVDbSys1
mount /dev/mapper/VGExaDb-LVDbSys2 /mnt/LVDbSys2

vi   /mnt/LVDbSys1/var/log/cellos/dbnodeupdate.log
vi   /mnt/LVDbSys2/var/log/cellos/dbnodeupdate.log

go to end of the line - compare the time stamp. which is latest.

e2label /dev/mapper/VGExaDb-LVDbSys1 - check the label
e2label /dev/mapper/VGExaDb-LVDbSys2

ls -al /mnt/LVDbSys2

e2label /dev/mapper/VGExaDb-LVDbSys2 ''
e2label /dev/mapper/VGExaDb-LVDbSys1 DBSYS

e2label /dev/mapper/VGExaDb-LVDbSys1- check the label
e2label /dev/mapper/VGExaDb-LVDbSys2

mkdir /mnt/boot
ls -lrt /mnt/boot
mount /dev/sda1 /mnt/boot
df -h
tar cvf /mnt/LVDbSys1/failed_boot.tar /mnt/boot
ls -lrt /mnt/LVDbSys1

rm -rf /mnt/boot/*
tar -pjxvf /mnt/LVDbSys1/boot_backup.tbz -C /mnt
echo "(hd0) /dev/sda" > /tmp/grubmap

/sbin/grub --device-map=/tmp/grubmap

find /I_am_hd_boot

root (hd0,0)

setup (hd0)

quit

sync

umount /mnt/LVDbSys1

umount /mnt/LVDbSys2

umount /mnt/boot
disconnect the CD-ROM image from the Devices menu