I am attempting to upgrade our OEM 11g installation to OEM 12c using the 2-System Method. I have completed the following steps -
1) Applied patch 13943512, Enterprise Manager 12c Upgrade Console to the 11g OEM, configured plugins and it seems to be functioning properly.
2) Created a copy of the 11g repository on a new server. The existing repository is 220.127.116.11.0 on Solaris Sparc. I installed 18.104.22.168.0 and patched to 22.214.171.124.0 on linux enterprise 5 x86-64 (I am told the repository must be the same RDBMS version in order to upgrade) and used export/import to move the repository to the new server.
3) On a linux 6 x86-64 I have downloaded the files -
as well as the plugins and am attempting to install the OEM. I executed the following command -
./runInstaller -pluginLocation /usr/local/oracle/oem12cagentplugins
I specified the new repository information at the connection details page and the following error is displayed -
Ensure that all EM AQ tables are enabled for enqueue and dequeue
I have searched the documentation looking for information on this requirement and am unable to find anything. I don't understand why it is requiring this and am looking for help in answering/resolving this question.
What steps have you followed to do the export/import? I believe it is not officially supported to do export/import for repository upgrade using 2-System Method. However, I was able to do the upgrade from 10.2.0.5 to 126.96.36.199.0 by following note this note "Grid Control Repository: Steps for Cross Platform Migration of the 10g Grid Control Repository [ID 388090.1]". See note "11g Grid Control: Steps for Migrating the 11g Grid Control Repository from One Database to Another [ID 1302281.1]" for 11g steps. There are post-installation steps to re-enable the queues.
I recommend to open an SR for help with repository cross-platform migration for upgrade 12c upgrade.
Thank you both for your replies. Nelson, I am not aware of the two notes you referenced and will take a look at them. I did open an SR with Oracle and was told that cross-platform migration between the version of Solaris SPARC and Red Hat Eneterpirse Linux is not supported and that I would thus have to do it using datapump export/import. I will check out the two notes you suggested. Hopefully the procedure I followed will not cause further problems as I continue.
I did figure out that the problem resulted from enqueue status of several jobs in the 11g database was set to yes but that status was not carried over when I restored the database on my new repository server. Once I did this -
exec dbms_aqadm.start_queue(queue_name => 'SYSMAN.<queue>');
for each of the the installation proceeded so it turns out that was the immediate problem. I also had to change some init parameters in the restored database. It would have been much cleaner if I could have used RMAN.The installation is now proceeding and I will continue on until the next hurdle is reached.