It has been an exciting week, Oracle 12c R2 came out and suddenly was time to refresh the RAC test environments. My friend Jacques opted for an upgrade from 184.108.40.206 to 220.127.116.11 (here the link to his blog post), I started with a fresh installation, because I also upgraded the Operating System to OEL 7.3.
Compared to 12c R1 there are new options on the installation process, but general speaking the wizard is quite similar.
The first breakthrough is about the installation simplified with an image based, no more runIstaller.sh to invoke but …
Unpack the .Zip file directly inside the Grid Infrastructure Home of the first cluster node as described below:
[grid@oel7node06 ~]$ mkdir -p /u01/app/18.104.22.168/grid [grid@oel7node06 ~]$ chown grid:oinstall /u01/app/22.214.171.124/grid [grid@oel7node06 ~]$ cd /u01/app/126.96.36.199/grid [grid@oel7node06 grid]$ unzip -q download_location/grid_home_image.zip # From an X session invoke the Grid Infrastructure wizard: [grid@oel7node06 grid]$ ./gridSetup.sh
The second screenshot list the new Cluster typoligies available on 12c R2:
- Oracle Standalone Cluster
- Oracle Cluster Domain
- Oracle Domain Services Cluster
- Oracle Member Clusters
- Oracle Member Cluster for Oracle Database
- Oracle Member Cluster for Applications
In my case I’m installing an Oracle Standalone Cluster
And now time for testing.
7 thoughts on “Installing Oracle Grid Infrastructure 12c R2”
Thank you for sharing this.
You checked the “Configure Oracle ASM Filter Driver” option on the ASM configuration page, so you’re not using ASMlib, right?
You had to configure the disks beforehand using asmcmd, right?
Yes I used ASM Filter Driver, I do not use ASMib and the disks or LUNs have to be configured before installing the GI.
I didn’t use ASMLIB, I’ve used NAS, iSCSI target devices with 12c Release 2 on Red Hat Linux 7.3
LikeLiked by 1 person
From your screenshot, I saw you installed Grid even you had several failures in the verification results: Soft limit, Port Availability, resolve.conf Integrity, and DNS/NIS name service. Isn’t that you need to fix all “failed” before proceeding? Also, it looks like you had two failures during the installation “Configure Oracle Grid Infrastructure for a cluster” and “Oracle Cluster verification Utility”…that doesn’t look like a successful installation to me…
I absolutely agree with you, in a real customer environment all the problems listed on the Prerequisites Checks window should be fixed before proceeding.
In my blog post, exception made for the Soft Limit which I fixed at the end of the installation, not all prerequisites were available (for example no DNS), because I used a standalone lab environment. Therefore, also the CVU execution at the end of the installation failed, because the DNS was not properly configured and the SCAN had only one thread. But for a first evaluation this setup was fine.
Thanks for the clarification Emiliano and also thanks for sharing your knowledge!