Upgrading Enterprise Manager Cloud Control to

Enterprise Manager Cloud Control is released yesterday, and I have downloaded it immediately to upgrade our system. It was a very smooth process. The version of our Enterprise Manager was Enterprise Manager is compatible with agents, so total downtime was about 30 minutes (of course it depends on your hardware). I upgraded our system using “one system upgrade” option and then upgraded all EM agents. Only a few agents are failed to upgrade, and I remember that we also had problems when deploying those agents. In this blog post, I will go through the whole process step by step.

I recommend you to read the official upgrade documentation (Thanks to Brandon to remind me to write this). As a blogger, I share my own experience. I can not guarantee that these steps will work on your environment, so you should always check the official documentation first.

As usual, you can download the new version from Oracle website. You need to download all 3 zip files to your server, create a directory and unzip them all into this directory:

Before you run the installer, we need to prepare our system for upgrading. Run emctl to copy EMKey from emkey.ora file to the management repository database:

Enter weblogic admin password as Admin User’s password and SYSMAN password for EM root password. The EMKey will be copied to the Management Repository (it will be removed automatically by the installer after the upgrade is done). Now shutdown the OMS:

Enterprise Manager will check if SYSMAN and DBSNMP has execute privilege to DBMS_RANDOM package and PUBLIC role has NO access to DBMS_RANDOM. So connect to your repository database as SYS and issue the following commands:

You also need to be sure that there are no invalid packages on repository database:

Although there were no invalid objects belong to these users, Enterprise Manager warned me about invalid objects. Maybe it doesn’t like to see that I have invalid objects in my APEX schema. I just ignored its warning.

An important warning: You should surely back up your repository database before starting to the upgrade process!

If there are no invalid objects in EM repository database and you had a valid backup, we are ready to go. Just run the installer:

Oracle Support Details

You enter Oracle support email and password or uncheck “receiving security updates” and click next.


It’s released yesterday so I just skipped checking software updates.


Verify that all prerequisite checks are succeeded. Click next.


I have chosen the one system upgrade. As you can see it detects the correct home of the existing Enterprise Manager installation.


Enter the new middleware home for upgraded Enterprise Manager, and click next to continue.


Enter the SYS and SYSMAN password of the repository database. As you can see, the installer reminds you to backup your repository database. Check if you have already backup your database, and click continue. At this step, the installer also warns you about required patches, suspending jobs during upgrade. Click OK and YES to these dialogs and go to next step.


Enterprise Manager will show a summary about plugins. As you can see it will update some plugins, keep some and also deploy some mandatory plugins. Click next to continue.


In this step, you can add additional plugins. Check the ones you want and click next.


Enter the weblogic admin password of the existing weblogic installation, and the path of OMS base location. It’s recommended to enter a path under your new middleware!


Review the inputs and click install if everything looks OK.

Almost done

It took about 30 minutes on our site. After the installation is completed, you need to login as root and run a shell script. After running the script, click OK to finish the installation.

Upgrade Agents

Now all you need is to upgrade your agents. Login to Enterprise Manager, open “upgrade agents” located under “manage cloud control” in setup menu. Add the agents to the list, and click submit.

Ignore the warning

If you haven’t set privileged credentials, or you don’t want to enter this info right now, you can just ignore the warning. That’s all. As I said, it was a very smooth process and everything went fine.

Please share

AWS Big Data Specialist. Oracle Certified Professional (OCP) for EBS R12, Oracle 10g and 11g. Co-author of "Expert Oracle Enterprise Manager 12c" book published by Apress. Awarded as Oracle ACE (in 2011) and Oracle ACE Director (in 2016) for the continuous contributions to the Oracle users community. Founding member, and vice president of Turkish Oracle User Group (TROUG). Presented at various international conferences including Oracle Open World.


  1. Courtney

    Just out of curiosity, how big is your site?   Upgrade timing can increase on larger sites so I’m trying to get a baseline for other customers 😉

    • Gokhan Atil

      It’s a small site, containing less than 50 hosts and 300 targets. EM12c OMS and Repository Database are running on same host.

  2. Great news. I’m hoping my upgrade goes as well. I’ll probably try an upgrade at home on the weekend. If that goes well, pencil in one for work soon…



    • Gokhan Atil

      Brandon, thanks for your comment. As you see, it has completed without any problem (nothing ruined), but you’re absolutely right that I should have mentioned the documentation. Although I checked the documentation, I forgot to write it. I’ll add it to post. On the other hand, most of them are not related my system (such as database service instance creation requests, snapshots of EM tables, BI Publisher etc ) and some of them are already implemented (because I have an uptodate repository database

  3. Tony Price

    So EM12c is certified with database repository.

    Is anyone considering upgrading repository database to with this upgrade?

  4. Kenan

    Merhaba Gökhan Hocam ben linux ĂŒzerinde’den upgrade yaptım’e.

    Daha sonra linux ĂŒzerinde ki agent’leri de upgrade yaptım fakat windows agent’ler olarak gözĂŒkĂŒyor ve listede windows agent yok.Windows agent dosyası var mı,varsa nereden indirebilirim bu dosyayı?ƞimdiden TeƟekĂŒrler

    • Gokhan Atil

      Selam Kenan,

      HenĂŒz windows için agentları yayınlanmadı. Bir iki haftaya yayınlanır diye dĂŒĆŸĂŒnĂŒyorum.

  5. Pingback: Enterprise Manager – Post Upgrade Tasks | Maaz Anjum's Blog

  6. samidee

    Hi, thanks for the article,

    is it possible to uninstall old middleware home after succefull install?

    thanks and regards

  7. Thank you for this. I’m in the planning stages, using this documentation as well as the official documentation from Oracle.

    In my installation I see these invalid synonyms, the comment for each one says ‘deprecated’. I don’t see a reference to them anywhere, if it’s okay to continue. I’m digging deeper to figure out why they are invalid, but perhaps I don’t need to.

    owner object_name

  8. Okay, so forget that last comment, sorry. It was one of those quite confusing synonym things where I had to access the object to make it VALID. Sometimes a DESCRIBE worked, other times a SELECT * was needed.
    I hate that.

  9. Hi,

    Maybe someone here can help:

    I try to upgrade EM to EM

    When the Repository-Upgrade is fails and in the Logfile I can not see anything
    about what’s going wrong.

    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Still Running…
    INFO: oracle.sysman.top.oms:Completed…
    INFO: oracle.sysman.top.oms:Starting both repos upgrade and transx loading in parallel
    INFO: oracle.sysman.top.oms:Running the repository in the upgrade mode
    INFO: oracle.sysman.top.oms:The command run for transx is /opt/oracle/product/EM_12.1.0.4/oms/sysman/admin/emdrep/bin/RepManager -connect (DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=podb2.lan.hbd)(PORT=1521)))(CONNECT_DATA=(SID=pdboem2))) -action transx  -repos_user SYSMAN -dbUser SYS -mwHome /opt/oracle/product/EM_12.1.0.4 -mwOraHome /opt/oracle/product/EM_12.1.0.4/oms
    INFO: oracle.sysman.top.oms:/bin/sh: module: line 1: syntax error: unexpected end of file
    INFO: oracle.sysman.top.oms:/bin/sh: error importing function definition for `module’
    INFO: oracle.sysman.top.oms:processing arguments
    INFO: oracle.sysman.top.oms:compiling arguments for validation
    INFO: oracle.sysman.top.oms:stty: standard input: Invalid argument
    INFO: oracle.sysman.top.oms:stty: standard input: Invalid argument
    INFO: oracle.sysman.top.oms: Enter SYSMAN user password   :
    INFO: oracle.sysman.top.oms:stty: standard input: Invalid argument
    INFO: oracle.sysman.top.oms:stty: standard input: Invalid argument
    INFO: oracle.sysman.top.oms:Verify :
    INFO: oracle.sysman.top.oms:Enter password for: SYSMAN
    INFO: oracle.sysman.top.oms: process_id:1dnhqxd3bexuq
    INFO: oracle.sysman.top.oms:[Initializing the EM RCU Runner, Running the Schema Action, Schema Action Completed Successfully]
    INFO: oracle.sysman.top.oms:The transx command execution is completed with the status 0
    INFO: oracle.sysman.top.oms:process_id:
    INFO: oracle.sysman.top.oms:The plug-in Repository Upgrade has failed its perform method

    I tried also the RepManager command on command line which succeeded and gave exactly the same output.

    I am lost somehow since I can not see any error message I can work with anywhere :/

    Any idea what’s going wrong here?

    best regards,


    • Fabrice

      Hello Rainer, any update on your problem?
      I’m facing exactly the same…


  10. Steve S.

    Thanks for the post. One thing to note is the upgrade document recommends installing the gc_inst outside of the middleware home like in previous 12c versions.


  11. Hemant Patel

    My installer was hung at 62% with below messages in the log as well. I checked the database alert log and trace files. I found the Oracle document “Detailed Steps for Troubleshooting ORA-00600 [kdsgrp1] (Doc ID 1492150.1)”.  The “Troubleshooting Steps” 1 and 2 were helpful in resolving my issue. OEM was successfully upgraded from to EM HTH.

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

    INFO: oracle.sysman.top.oms:Still Running

  12. Stephane

    Hi Atil,

    Would you please let know if you have the same issue regarding CRC or md5sum check?


    Oracle Enterprise Manager Cloud Control 12c Release 4 (
    for Linux x86-64 :
    -Download em12104_linux64_disk1.zip (2,195,693,096 bytes) (cksum – 4176170445) cksum OK

    -Download em12104_linux64_disk2.zip (1,877,449,643 bytes) (cksum – 3219437597) cksum is different :1825566964

    Download em12104_linux64_disk3.zip (2,727,123,784 bytes) (cksum – 2554054536) cksum OK

    I re-downloaded several time em12104_linux64_disk2.zip and also from Edelivery having same problem!


    • Gokhan Atil

      Hi Stephane,

      Here’s my checksum result:

      cksum em12104_linux64_disk2.zip

    • Stephane

      Thank you for the confirmation. I am having
      a weird problem with this zip file.

      cheer up for your blog!

  13. Stephane

    Hi Atil,


    I have open a proactive SR with MOS (aka Metalink) regarding my OEM 12 R1.

    I am weighing the Pros and Cons : upgrade or fresh install with 12c R4 (and patch).

    here Oracle support feedback, what do you think about it?

    As discussed earlier please check the following 2 approaches we can try 1) The one currently we are working on + Apply BP1 patches on OMS + Apply BP1 patches on all the agents + Upgrade OMS to + Upgrade all agents to Advantages >> This approch will ensure all the historical data about the targets monitored is intact >> All the created jobs users etc will be carried forward to new version Disadvantage >> Takes lot of time to apply BP1 on all the agents >> Needs downtime and incase of failure need to restore back to start again 2) To directly go to Release + Install OMS on a new host + Install agents in parallel with agent ( on all the targets + Once everything is up and all targets are correctly monitored using + Stop Advantages >> No down time required, the monitoring will be going on inn setup which is untouched >> You can roll back safely without affecting the setup Disadvantages >> You will loose all the historical data >> All the jobs, users, notification rules need to be created manually 

    I am looking to have a second opinion, please let me know what would be your approach ?


    Stephane Berhault

    • Gokhan Atil

      Stephane, it depends on you. If historical data is important for you, if you have set up lots of job, incident rules etc, you may prefer upgrade path. Applying BP1 patches could be really challenging (and surely time consuming).

      I don’t need to keep historical data (disk usage stats are very useful but I can live without them). We have less than 100 targets, 2-3 jobs and not-well customized incident handling rules. So I would surely go with a fresh installation for my system.

    • Stephane


      Thank you for your quick answer. Like you said applying BP1 is quite challenging. My colleagues in France did it and lost their entire OEM after BP1 failure.

      I am going to do a fresh install with OEM 12c Release 4.
      Have you already apply the patch?

      Patch 18945232: ENTERPRISE MANAGER

      take care

  14. sachin

    I used your procedure and it went pretty smooth.

    it is now upgraded to 12cR4


  15. sachin

    All went good but $OMS_HOME/bin/emctl is not working because that file is just went to zero size.

    -rwxr–r– 1 oracle dba     0 Aug  5 13:51 ./oms/bin/emctl

    How does that happen? how do I get that thing back?


    • Gokhan Atil

      Sachin, I don’t know why it happened. You may check the logs of upgrade process. Please also check: EM12c Recreating the emctl Script for the Enterprise Manager Cloud Control OMS (Doc ID 1404623.1). It may help.

    • sachin

      Thanks for reply,
      I check that document and nowhere I can find how to recreate emctl

      I just copy over from 11g installation but it shows old version
      >$OMS_HOME/bin/emctl status oms
      Oracle Enterprise Manager Cloud Control 12c Release 3

  16. sachin

    Man oh man I can’t find anywhere on how to get emctl.

    so on my 2gig linux desktop I install 12cR4 and copy that emctl and it is working now.

    backed up whole Middleware installation for safe keeping as it is not that big.

  17. sunil

    Is this out of Place upgrade ?  or  Does OMS goes in to the same existing OMS home

  18. George

    Excellent documentation Gokhan. It’s help me a lot to upgrade my OEM …


  19. Zhuang Li

    Hi Gokhan,

    I am at the planning stage of doing this kind of upgrading.

    Read you poster and doc, a few questions:

    1) Do i need to stop all agents running before upgrading?

    2) Perhaps there is no need to upgrade that agents if it is as they are compatiable with

    3) My repository is on version, according to the doc, I should run a patch, but not compulsory. When check the patch, the readme says, if not the kind of error, not run. Guess I shouldn’t run, yes?

    Many thanks,



    • Gokhan Atil

      Hi Zhuang,

      Here are the answers:

      1) No you don’t.
      2) As you said they will work without upgrading but I highly recommend you to update the agents. By the way, it’s a very easy task, so why don’t you want to use old agents?
      3) Well, the upgrade process may be OK without patching but I prefer to apply the latest patchs to my DB as long as it’s compatible with the application. So I recommend you to patch it 🙂

  20. Kumar

    Gokhan ,

    Thanks for sharing valuable info in detail .  We  are also in the process to upgrade to from . I have questions

    We have 2 OMS servers running on 2 diffrent hosts . on hosts1 we have Admin , OMS1 , JVMD Engine Server,ADP Engine servers are running and on Second host we have OMS2 is running . both host have thier own installs ( I mean Path is same in both hosts,but not shared FS) and bot have their own OHS instances .  What method do  I need to follow to upgrade it ? One system upgrade   on both hosts or 2 -system upgrade on both  hosts   ?

    • Gokhan Atil

      Kumar, it depends on how much downtime your system can tolerate. If you have resources (disk, CPU, RAM) on the server, you can use two system approach. Otherwise, use one system upgrade.

  21. Zhuang Li

    took nearly 3 hours to finish. at one point, it failed. checked Oracle Support and followed:

    M 12c: Plugin Upgrade Failed on OMS With Error Server-side plug-in deployment failed. Error in schema configuration: ORA-20778 (Doc ID 1912238.1)

    then retry and carried on. Finally it finished 🙂

  22. Monte

    Hi Gokhan,

    Thanks for the information! I am performing the EM CC –> upgrade on two separate OMS hosts. The 1st has the Admin Server on it. Both are serving the same EM Console end-point in a load balanced environment. The upgrade finished successfully on the 1st host. However, it fails on the 2nd at 80% during the ‘Start Oracle Management Service’ step. This occurs after the ‘OCM Configuration for OMS’ step successfully completes and the WebTier successfully starts. The referenced EMGC_OMS2.out file shows the following error;

    “The server name EMGC_OMS2 specified with -Dweblogic.Name does not exist. The configuration includes the following servers {EMGC_ADMINSERVER}”

    I noticed that the config.xml on the 1st host has the EMGC_OMS2 entry, but the config.xml on the 2nd host doesn’t. Oracle says that is to be expected.

    Oracle Support seems to think it might have something to do with either the two OMS nodes not being able to communicate, or possibly having something to do with a certificate issue? However, they seem as lost as I do right now.

    The documentation says;

    “If you have done any of the following customization to the OMS, then ensure that you remove all of them before starting the upgrade process. Once the upgrade is complete, you can redo the customization

          Additional third-party SSL certificate.”

    We have a 3rd party certificate configured for our EM console and Admin server console, but Oracle says that shouldn’t cause an issue during the upgrade.

    Any ideas?

    • Monte

      Just wanted to update that this has been resolved with an opening of port 7403 b/w the two OMS hosts and a start of the OMS server (EMGC_OMS2) on the 2nd OMS host (which was the one that failed on the OMS startup step in upgrade) via EM Admin console. According to Oracle, this somehow ‘resynched’ the nodemanager files.

      Thanks, Monte

    • Adil

      Hi Gokhan,
      I got the same error after EMGC_OMS2 crash and re-install.
      The omsca recover failed at the configure OMS step : emctl start oms -admin.

      When I checked the pluginca logs , I observed that the process omsca recover get an old OMS2 GUID ” BD02A41D0B0CB234DD2FDEA340862150″

      And in the end I got the error below :
      [main] FATAL swlib.storage – Security context instance is null!
      2015-03-26 17: 07: 13.556 [main] ERROR pluginca.PluginCADriver – Error while Registering older plugins.

      Any ideas ?

      Thank you very much !

  23. Raghu

    Hi Gokhan,

    We have two OMS servers ( which are in sync and repository database is running on two node RAC and its version is

    The following versions are existed ones:

    OMS Version :

    Plugin: Database Plugin :

    OMS Patch: 18604893 and 13983293

    I need to upgrade it to :

    OMS Version :

    Plugin: Database Plugin :

    OMS Patch: 19218456

    My doubt is: Can’t we upgrade OMSes from to through EM12c GUI, so that ‘patch & provision’ approach will take care both OMSes in one go? I would like to avoid manual way (if possible).


    Thanks for your help.




  24. hello Gokhan Atil,

    just the same procedure for upgrade you did apply for upgrading grid 11g to 12c cloud.if not do you have a step by step documentation for that.


  25. Ravindra malwal


    Nice Document sir, i will do it first my home…. 🙂

  26. Sam

    Hi Gokhan,

    I have recently started following your website and you have done an amazing job in sharing your knowledge.

    I would really appreciate if you have any information for OEM upgrade from to

    Like the steps you gave for Your help is much appreciated.


    • Sam

      Hi Gokhan,

      Thank you for the info. I tried to follow your upgrade for our approach.

      However from my Installer in the Installation Type:

      For the Upgrade an existing Enterprise Manager System

      And in this when I tried to select “One System Upgrade”, I’m not able to select this option, for some reason it is greyed out.

      Any help on this is really appreciated.

  27. Tan Do

    We upgrade from to 12.1.04 and get these error.  Also see Oracle fill a bug for this without solution however.  Anyone face this problem let me know.




    Here is the error we received:

    INFO: oracle.sysman.top.oms:    at oracle.sysman.core.common.extensibility.metadata.EMMetadataServiceInfo.parseMdInstance(EMMetadataServiceInfo.java:851)
    INFO: oracle.sysman.top.oms:    … 4 more
    INFO: oracle.sysman.top.oms:Caused by: oracle.sysman.emInternalSDK.core.selfupdate.integration.OperationFailedException: ORA-20658: Invalid status – 5. Please refresh and try again.
    INFO: oracle.sysman.top.oms:ORA-06512: at “SYSMAN.EM_SELF_UPDATE”, line 536
    INFO: oracle.sysman.top.oms:ORA-06512: at “SYSMAN.EM_SELF_UPDATE”, line 350
    INFO: oracle.sysman.top.oms:ORA-06512: at line 1
    INFO: oracle.sysman.top.oms:
    INFO: oracle.sysman.top.oms:    at oracle.sysman.emInternalSDK.core.selfupdate.InventoryManager.registerOutOfBoxEntity(InventoryManager.java:847)
    INFO: oracle.sysman.top.oms:    at oracle.sysman.core.selfupdate.EntityInstRegistration.parse(EntityInstRegistration.java:185)
    INFO: oracle.sysman.top.oms:    … 5 more

  28. Patricio Mera

    Hi Gokhan, before updating to I  am patching a agent for AIX systems; I want to know if  the message “Shut down All Nodes, apply patches, startup, apply SQL (if needes), downtime required.” showed during deployment refers to hosts, database or agents?

    • Gokhan Atil

      Patricio, if you’re patching Enterprise Manager agents, it should refer to the EM agents.

  29. Kamal

    Hi Gokhan,

    Nice document and i was able to upgrade OMS & agent from to successfully.

    Thank you.


    • Gokhan Atil

      Hi Kamal, I’m glad to hear that! You’re welcome 🙂

  30. fusion


    can we install older one after successful installation of newest version can you explain in detail.

    • Gokhan Atil

      Theoretically yes you can but you should use different ports for each installation. It will surely mess all things 🙂 So do not!

Leave Comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.