Richmond Cluster (06)

Saturday 03 May, 2008 - 22:56

Set up the ssh configuration file (on richmond1 and richmond2 ) to ensure that X11 forwarding will not cause the installation to fail (p.2-14 of Clusterware Install for Linux ):
$ su - oracle
$ cd ~/.ssh
$ cat >config
Host *
ForwardX11 no

Setting the raw devices according to the design in Richmond Cluster (02) by appending the configuration information to /etc/sysconfig/rawdevices , restarting the service, and checking the status. This is done on both richmond1 and richmond2 :
$ su -
# cd /etc/sysconfig
# cat >>rawdevices
/dev/raw/raw1 8 17
/dev/raw/raw2 8 18
/dev/raw/raw16 8 33
/dev/raw/raw17 8 34
/dev/raw/raw31 8 49
/dev/raw/raw32 8 50

# service rawdevices restart
# raw -qa
/dev/raw/raw1: bound to major 8, minor 17
/dev/raw/raw2: bound to major 8, minor 18
/dev/raw/raw16: bound to major 8, minor 33
/dev/raw/raw17: bound to major 8, minor 34
/dev/raw/raw31: bound to major 8, minor 49
/dev/raw/raw32: bound to major 8, minor 50

Created the home directory for clusterware (see pp.2-43 to 2-44 of Clusterware Install for Linux ) on both richmond1 and richmond2 :
$ su -
# mkdir -p /u00/app/oracle
# chown -R oracle:oinstall /u00/app/oracle
# chmod -R 775 /u00/app/oracle
# mkdir -p /u00/crs/oracle/product/10/app
# chown -R root:oinstall /u00/crs
# chmod -R 775 /u00/crs

Ready to start the clusterware installation on richmond1 :
$ su - oracle
cd ~/installation/clusterware/
$ export ORACLE_BASE=/u00/app/oracle
$ export ORACLE_HOME=/u00/crs/oracle/product/10/app
$ umask
0022
$ xterm &

Installed the hangcheck_timer (see pp.2-46 to 2-47 of Clusterware Install for Linux ) on both richmond1 and richmond2 :
$ su -
# lsmod|grep hang
# insmod hangcheck-timer hangcheck_tick=30 hangcheck_margin=180
Using /lib/modules/2.4.21-40.EL/kernel/drivers/char/hangcheck-timer.o
# lsmod|grep hang
hangcheck-timer 2616 0 (unused)
# cat >>/etc/rc.d/rc.local
/sbin/insmod hangcheck-timer hangcheck_tick=30 hangcheck_margin=180

I needed to redo the disk partitioning because I got the sizes wrong according to p.3-10 of Clusterware Install for Linux . On richmond2 only:
$ su -
# service rawdevices stop
# parted /dev/sdb
(parted) rm 1
(parted) rm 2
(parted) mkpart primary 1 257
(parted) mkpart primary 257 513
(parted) select /dev/sdc
...
(parted) select /dev/sdd
...
(parted) quit
# service rawdevices start

I restarted the rawdevices service on richmond1 .

Needed to set up the correct ownership and security on the raw disks (see p.3-15 ibid. ) on both richmond1 and richmond2 :
$ su -
# service rawdevices stop
# chown root:oinstall /dev/raw/raw*
# chmod 640 /dev/raw/raw*
# service rawdevices start