How to : Clean deinstall of Oracle RAC (CRS since 11.2.0.1) –> On Linux Red Hat 5
First check your CRS Version to see if this doc is applicable:
crsctl query crs activeversion
If CRS version is 11.2, it’s ok to start 😉 :
–> Stop all CRS on the two nodes.
with oracle account or root (I recommand to use root account)
crsctl stop crs -f
or (root account only)
/etc/init.d/ohas stop
-> Check that no more processes are running
Then, delete the Oracle_base directory and Oracle_GRID directory. I don’t make it with « deinstall » script, because some files or directory always exists and the reinstall of complete CRS failed.
Cleanup inittab, remove :
"h1:3:respawn:/sbin/init.d/init.ohasd run >/dev/null 2>&1 </dev/null"
==> cleaning (delete) the files / directory bellow
rm -rf /etc/oracle
rm -rf /var/tmp/.oracle
rm -rf /tmp/.oracle
rm -rf /var/opt/oracle
Launch this command to delete the rc* file belong to ohas :
find /etc/rc*.d -name *ohasd -exec rm {} \;
==> Check your ASM Device and delete the « marked » files
Example :
blkid /dev/mapper/orap1
/dev/mapper/orap1: LABEL="" TYPE="oracleasm"
Launch the command bellow : (Warning : this command delete all of DATA in the ASM drives !!!)
dd if=/dev/zero of=/dev/mapper/orap1 bs=1024k count=100
check again the mark of the file
blkid /dev/mapper/orap1
Delete other files
rm /etc/oratab
rm -f /usr/local/bin/dbhome /usr/local/bin/coraenv /usr/local/bin/oraenv
Now, you can start the complete installation of Oracle Grid infrastructure 11.2
Thanks to « Uggla » Blog (http://uggla.free.fr/wordpress) and the help of Arkzoyd (http://blog.arkzoyd.com), that allow me to make a clean deinstall of my cluster 😉