In order to upgrade our OEM environment from 220.127.116.11 to 18.104.22.168 we need to download the latest version of OEM from either http://otn.oracle.com or from http://edelivery.oracle.com. Once we have the files needed to upgrade, we need to stage them in a location that is central.
Note: Most downloads on Linux will be placed in the /home/oracle/Downloads directory.
Once we are in the directory where the downloaded files are, we need to unzip these files.
#> unzip em12103_linux64_disk1.zip –d ./22.214.171.124
The zip command needs to be repeated for all of the files downloaded.
Next, we need to move into the directory where we unzipped everything and run the runInstaller.
#> cd ./126.96.36.199
#> ./runInstaller &
When the installer starts, we will be greeted with our standard Oracle Installation Wizard. To upgrade the OEM to 188.8.131.52, consists of nine steps. We will review these nine steps throughout this posting.
When asked if we want to be notified of security issues, we need to uncheck the box. Say “yes” to whatever prompt is presented when clicking next.
We want to skip any updates for this version. Since this version is the current release, there is no additional items to be added. Click Next.
Make sure that all the prerequisites are meet. In the screen shot, we see that I have a warning. This warning can be ignored; however, later on we will see a make error, which can be ignored as well.
On this step, we tell the installer to do an upgrade. In the screen shot, we are doing a 1-system upgrade.
Part of the upgrade, Oracle will be doing an out-of-place upgrade. This means we need to provide a new Oracle Home directory for the middleware stack.
Provide the password for the repository database. These passwords are normally the SYS and SYSMAN passwords. This upgrade process will be modifying the SYSMAN tables as needed.
The upgrade process will be looking for the “emkey“. This key needs to be placed inside the database so it can be referenced during the upgrade. If not available, the wizard will error and provide you with this message.
To prevent this message for triggering, run the following command before the upgrade is started. If already started, we can still run this command and move pass the error message.
#>./emctl config emkey –copy_to_repos –sysman_pwd <password>
Additional errors may arise. This is due to the database not having all the correct parameters set at the instance level. The parameters identified in the error message can be ignored. Click “Yes” or “OK”.
During this step, Oracle is telling us what versions of the plug-ins we currently have installed and what version will be available after the upgrade is done. Click “Next”
If we need to add additional plug-ins, we can at this time. For demo purpose of this upgrade, we don’t need to add any additional plug-ins.
Before we can do anything to the Weblogic layer, we need to bring down the OMS. If the weblogic server is up the wizard will complain and tell us to bring it down.
#> cd /opt/oracle/middleware/oms/bin
#> ./emctl stop oms –all
The wizard will identify the correct Weblogic server and port number. The only thing we need to provide is the password for the weblogic user. If the OMS Instance Base Location is not prepopulated, then fill this in as well. Click “Next”
We can review all the options that will take place during the upgrade.
This is the fun part. We get to watch Oracle Enterprise Manager upgrade itself. Depending on the speed of the box, it may take a bit to upgrade.
The finish screen (not captured). On this screen it will tell you the URL for both the OMS and Admin server. With this being just an upgrade, everything remains the same. All previous URLs should work (illustrated below).
Once everything is up and accessible. We can check the version of OEM from the web interface.