Database Reference
In-Depth Information
Listing 11-9.
Additional Configuration to the Requirements Listed on MOS Note 1330701.1
[root@oevm1 ~]#
yum upgrade
[root@oevm1 ~]#
yum install portmap
[root@oevm1 ~]#
yum install oracle-rdbms-server-11gR2-preinstall
[root@oevm1 ~]#
chkconfig rpcbind on
[root@oevm1 ~]#
chkconfig nfs on
[root@oevm1 ~]#
chkconfig iptables off
[root@oevm1 ~]#
service rpcbind start
[root@oevm1 ~]#
service nfs start
[root@oevm1 ~]#
service iptables stop
Use ACFS for the Application Tier Files
Using ACFS is the only way to configure a shared application tier file system without introducing external
dependencies. Virtualized ODA is configured with a single 50G ACFS volume, which is mounted on both ODA Base
VMs as
/cloudfs
. The 50G size is not sufficient for R12.2 implementations because the disk space requirement for
the e-Business Suite dual file system is 64G, so it has to be increased. My Oracle Support document “ODA (Oracle
Database Appliance): How To Resize CloudFS (Doc ID 1437717.1)” explains how to add and resize the ACFS volumes
on ODA. Listing 11-10 shows the steps to execute as the
grid
user on the ODA Base to increase the
/cloudfs
from 50G
to 150G.
Listing 11-10.
Resizing the /cloudfs
[grid@s1 ~]$
. oraenv
ORACLE_SID = [grid] ?
+ASM1
The Oracle base has been set to /u01/app/grid
[root@s1 ~]#
df -hP /cloudfs
Filesystem Size Used Avail Use% Mounted on
/dev/asm/acfsvol-352 50G 166M 50G 1% /cloudfs
[grid@s1 ~]$
/sbin/acfsutil size +102400m /cloudfs
acfsutil size: new file system size: 161061273600 (153600MB)
[grid@s1 ~]$
df -hP /cloudfs
Filesystem Size Used Avail Use% Mounted on
/dev/asm/acfsvol-352 150G 369M 150G 1% /cloudfs
Once you have resized the ACFS volume based on your needs, it's just a matter of mounting it to the VM server
using NFS. It might seem a straightforward task to add the volume to the ODA Base server in the
/etc/exports
file
and
/etc/fstab
file on the VM, which would allow you to mount the file system successfully, but you have to make
sure the mounted file system remains accessible, even if one of the ODA Base VMs goes down. Thus, a smarter
solution is required.
First of all, you need to make sure each ODA Base VM can act as the NFS server for any of the application tier
nodes (you have only one of them at the moment, but you know the IP address of the second one already). I'm adding
one entry for each application tier VM to both ODA Base nodes in /etc/exports, as displayed in Listing 11-11.
Listing 11-11.
/etc/exports on ODA Base VMs
[root@s1 ~]#
cat /etc/exports
/cloudfs 10.177.0.61(rw,sync,no_root_squash,fsid=1)
/cloudfs 10.177.0.62(rw,sync,no_root_squash,fsid=1)
