Oracle Fusion Applications Installation: Provision an Applications Environment

Previous: Create new Response File

Note: If you are using VMWare, it is advisable to keep the staging directory path to be short to avoid any file path going more than 255 characters.

[root@fahost ~]# umount /mnt/hgfs

[root@fahost ~]# mkdir /stage

[root@fahost ~]# mount -t vmhgfs .host:/stage /stage

Since we have copied required Webgate libraries on node 1, we must copy them on node 2 as well.

[fusion@fahost bin]$ mkdir /app/fusion/oam_lib/

[fusion@fahost bin]$ cd /app/fusion/oam_lib/

[fusion@fahost oam_lib]$ scp fusion@fmwhost:/app/fusion/oam_lib/* .

[fusion@fahost oam_lib]$ ls -ltr

total 1024

-rwxr-xr-x 1 fusion dba 58400 Apr 6 15:01 libgcc_s.so.1

-rwxr-xr-x 1 fusion dba 976312 Apr 6 15:01 libstdc++.so.6

[fusion@fahost oam_lib]$ cp -pr /app/fusion/oam_lib/* /stage/installers/webgate/

Following file is also required for provisioning phase in addition to above.

[fusion@fahost oam_lib]$ cp -p /usr/lib64/libstdc++.so.5 /app/fusion/oam_lib/

[fusion@fahost oam_lib]$ cp -p /usr/lib64/libstdc++.so.5 / stage/installers/webgate/

Make sure that the IDM infrastructure is started on Node 1. (only if not already started)

1. Source profile

more ~/.bash_profile

ORACLE_HOME=/app/fusion/database/product/11.2.0/dbhome_1

export ORACLE_HOME

ORACLE_SID=fusiondb

export ORACLE_SID

JAVA_HOME=/app/fusion/jdk6

export JAVA_HOME

PATH=$PATH:$ORACLE_HOME/bin

export PATH

2. Start Listener

lsnrctl start LISTENER_FUSIONDB

3. Start Database

sqlplus / as sysdba

SQL> startup

4. Start OID

/app/fusion/admin/oid1/bin/opmnctl startall

5. Start Web

/app/fusion/admin/web1/bin/opmnctl startall

6. Start Weblogic Admin Server for IDMDomain

cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/

nohup ./startWebLogic.sh &

(Wait till you see RUNNING in nohup.out log file)

7. Start Node Manager

cd /app/fusion/fmw/wlserver_10.3/server/bin/

nohup ./startNodeManager.sh &

(Wait till you see “started on port 5556″ in nohup.out log file)

8. Start OAM managed server for IDMDomain (since we cannot login to console without OAM)

cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/

nohup ./startManagedWebLogic.sh wls_oam1 &

(Wait till you see RUNNING in nohup.out log file)

9. Start remaining managed servers

Either from command prompt

cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/

nohup ./startManagedWebLogic.sh wls_ods1 &

nohup ./startManagedWebLogic.sh wls_oim1 &

nohup ./startManagedWebLogic.sh wls_soa1 &

(Wait till you see RUNNING in nohup.out log file for each of them. You can start them in in parallel, just make sure all start fine)

Or by logging into Weblogic Console->Servers->Control screen.

10. ONLY In case if restarting from postconfigure phase, do the following on Node 2 (FA node)

For 11.1.4 and earlier versions

cd /app/fusion/fusionapps/wlserver_10.3/common/nodemanager/ad002aph01

nohup ./startNodeManagerWrapper.sh &

For 11.1.5 onwards releases

cd /app/fusion/instance/nodemanager/ad002aph01

nohup ./startNodeManagerWrapper.sh &

-bash-3.2$ /app/fusion/instance/CommonDomain_webtier/bin/opmnctl startall

Start Fusion Applications Provisioning

Launch the installer from <provisioning_repository>/bin

It is important to start it with “-ignoreSysPrereqs true” switch as mentioned in the release notes.

[fusion@fahost bin]$ ./provisioningWizard.sh -ignoreSysPrereqs true &

Click Next

Select “Provision an Applications Environment” and browse for the response file created in last step. Click Next

Review and click Next

Validate the information and click Next

In case if you want to revisit some of the section in the Response File then you can do so here. Click Next

Review the summary and click Next to begin. Good luck J

You can ignore following error since we are not using OVD. Also since we have used “ignoreSysPrereqs true” you will see Next button. Click Next to continue.

Once the Installation phase finishes, click Next

Note: If you are using longer path for stage directory on Linux then install phase may fail with errors. This is due to more than 255 characters path for certain patch files. Check following link for solution for the same.

http://www.oratraining.com/blog/2013/04/solution-for-atgpf-build-failure-during-install-phase-with-oui-67073utilsession-failed-null-error/

Once Preconfigure phase completes, click Next

Important Notes:

You may see a lag between 100% completion of the phase and “Next” button being enabled. This is the time where the status changes from COMPLETED to ENDED. Here it takes backup of the instance directory at /app/fusion/provisioning/restart so that it can cleanup and restore from next step if it fails.

The installation logs are being created at/app/fusion/logs/provisioning/plan<timestamp>/<hostname> directory. This is the most useful place to troubleshoot.

The status flags like STARTED, COMPLETED, FAILED, ENDED are stored in/app/fusion/provisioning/phaseguards directory. Please be careful while removing, renaming or creating flag files here since it will change the current status of the provisioning.

And for status of each domain within the particular phase, the flag files are written at/app/fusion/provisioning/phaseguards/pf directory

Good thing in new Fusion applications setup is that when you retry it only performs the actions for failed products but those actions of completed products are not repeated. This saves a lot of time compared to our last installation of 11.1.1.5

Once configure completes, click Next

Note: If configure fails with following error then do this workaround.

“Caused by: java.lang.RuntimeException: IO Error: Connection reset”

Workaround:

SQLNET.INBOUND_CONNECT_TIMEOUT = 0 in at $ORACLE_HOME/network/admin/sqlnet.ora on node 1 (or DB node)

LSNRCTL> set inbound_connect_timeout =0 on DB node

INBOUND_CONNECT_TIMEOUT_LISTENER_FUSIONDB = 0 in $ORACLE_HOME/network/admin/listener.ora on DB node

You can remove this once provisioning is complete.

Note: The name of next phase has changed to “Configure Primary/Secondary” instead of “Configure Secondary” as in 11.1.1.5 release since it was misleading.

Click Next once Configure Primary/Secondary phase completes.

Post configure is the most important and most error-prone part of Fusion Applications provisioning. Most of the people don’t get such a clean screen like above in first attempt (neither did I earlier). Don’t worry pr panic if you get errors, rest assured that if you have followed all earlier steps correctly, they are mostly random errors due to out of memory, network etc. make sure that you have sufficient free memory on node while running this. Also make sure that all components in IDM node (odi is optional at this phase) are up and running.

Validate that you are able to login to all IDM components before running this phase. Login to EM console to see the status: http://<fmwhost>:7777/em

Even if postconfigure has completed on some products and failed on others, when you click on retry, it will restore the failed products only and skip the phase for completed products.

You will see following screen when you click on Retry if any phase has errored out. Once you press OK it will restart the same phase only for failed product families.

Be patient when startup brings up all required managed servers. please note that if some components fail to startup and if you retry, it will do startup phase for all domains again including those have completed successfully. Make sure sufficient memory is available during this point.

Click Next once completed.

Validation phase has known bug and you will see following error for each domain validation.

During the Validate phase, you will encounter WebGate validation errors and the error messages in the provisioning log are as follows:

Webgate: private-validate validateWegPageStatus – BUILD_ERROR: The HTTP request to http://<host>:<port>/oberr.cgi?progid=1 returned status: 404

[runProvisioning-validate] [NOTIFICATION] [] [runProvisioning-validate] [tid: 11] [ecid: 0000JgMcCTD9lZOLIih8if1GeQ7k000002,0] [logStatus] STATE=BUILD_ERROR!TIMESTAMP=<> PST!TARGET=private-validate!CATEGORY=BUILD_ERROR!DOMAIN=CommonDomain!HOSTNAME=<host>!PRODUCTFAMILY=fs!PRODUCT=WebGate!TASK=validate WebPageStatus!TASKID=fs.WebGate.BUILD_ERROR.private-validate.validate WebPageStatus!MESSAGE=The HTTP request to http://<host>:<port>/oberr.cgi?progid=1 returned status: 404.!DETAIL=The HTTP request to http://<host>:<port>/oberr.cgi?progid=1 returned status: 404.!BUILDFILE=<framework_location>/provisioning/provisioning-build/webgate-build.xml!LINENUMBER=992!

These WebGate web page validation errors can be ignored. If there are any other validation errors you must resolve them before proceeding to the Summary phase. After resolving all validation errors, if the Next button on the Provisioning Wizard is not enabled, perform these steps from the command line to enable it:

  1. cd <APPLICATIONS_CONFIG>/phaseguards (for example /app/fusion/provisioning/phaseguards)
  2. rm validate-<host>-FAILED.grd.
  3. touch validate-<host>-COMPLETED.grd.
  4. touch validate-<host>-ENDED.grd (this last part is not mentioned in Oracle documents but this is required)
  5. The Next button should be enabled on the Provisioning Wizard.

WARNING:

Deleting and creating files in the phase guard directory should be used as a workaround to resolve validation phase issues ONLY if none of the other options work. In any other case, you should never modify or make changes to the phase guard files.

[fusion@fahost]$ cd /app/fusion/provisioning/phaseguards

[fusion@fahost phaseguards]$ rm validate-fahost-FAILED.grd

[fusion@fahost phaseguards]$ touch validate-fahost-COMPLETED.grd

[fusion@fahost phaseguards]$ touch validate-fahost-ENDED.grd

Now you can see the Next button enabled. Click Next to go to post-installation summary.

Summary:

ComonDomain

Admin Console: http://fahost:10633/console

EM Welcome page: http://fahost:10633/em

Homepage – Application URL: https://fahost:10634/homePage

CRM Domain

Admin Console: http://fahost:10635/console

EM: http://fahost:10635/em

Financial Domain

Admin Console: http://fahost:10623/console

EM: http://fahost:10623/em

HCM Domain

Admin Console: http://fahost:10639/console

EM: http://fahost:10639/em

BI Domain

Admin Console: http://fahost:10641/console

OIM

HTTP URL: http://fmwhost.paramlabs.com:7777

HTTPS URL: https://fmwhost.paramlabs.com:7777

Login to homepage URL with weblogic_fa user.

You can see that “Fusion Accounting Hub” is showing up as “Provisioned” but implementation is “Not Started” yet. This is the next step after installing Fusion Applications and not part of this guide.

Lot more coming soon..

Happy learning !

Tushar Thakker

Apr 14th, 2013 | Posted by tushar | In Uncategorized
  1. Sandesh
    May 15th, 2013 at 05:36 | #1

    Hi,
    We are doing two node installation. While running postconfigure stage of provisioning on Fusion Applications node we get the message:-
    “Waiting for taxonomy upload to complete on the CommonDomain host. You should ensure that postconfigure has been triggered on that host for this to proceed.”
    What does this signify? Does it mean we need to run postconfigure on fusion Middleware host also?

    Thanks

    • tushar
      May 15th, 2013 at 06:17 | #2

      Dear Sandesh,

      This message means that the postconfigure for common domain/Functional Setup (fs-postconfigure) is still not completed. Once fs-postconfigure passes taxonomy upload step which populates tables related to taxonomy, then automatically postconfigure for hcm etc will continue from that point. If fs postconfigure fails “before” finishing taxonomy upload stage, then these also will fail eventually after timing out.

      • Sandesh
        May 15th, 2013 at 06:32 | #3

        Hi,

        So do we need to manually trigger postconfigure on middleware host or it means installation is still in progress on FA host?

        Thanks
        Sandesh

        • tushar
          May 15th, 2013 at 06:39 | #4

          Dear Sandesh,

          Kindly let me know if you have selected one server for all domains or multiple servers for FA domains. If you have selected single node for FA domains then fs postconfigure would have already triggered and you can also see on the screen. Please let me know if you have different topology.

          Regards
          Tushar

          • Sandesh
            May 15th, 2013 at 06:45 | #5

            Hi Tushar,

            We have select one host for all domains. Once the above message was displayed, nothing was happening on the fa node. It was idle for long time.

            Thanks,
            Sandesh

          • tushar
            May 16th, 2013 at 08:39 | #6

            Was functional-setup(fs) postconfigure completed or ongoing? You can easily see it on the Screen if you used the wizard. The whole thing is related to fs-postconfigure completing taxonomy upload successfully. Unless this passes, hcm etc will keep waiting for it to complete and release lock file.

          • Sandesh
            May 17th, 2013 at 10:22 | #7

            Hi Tushar,
            functional-setup(fs) postconfigure was running. Does it take lot of time? Do we need to perform any step on fmw host?

            Thanks,
            Sandesh

          • tushar
            May 20th, 2013 at 04:38 | #8

            Dear Sandesh,

            Yes, postconfigure stage takes the highest amount of time among all stages. It may take many hours. also do not cancel the setup even if it is taking quote a long time since there are timeouts already present, so it will never go on forever. If it is running, it surely is doing something. sometimes the log files may not see any new activity for more than hour but that is normal.

            Thanks
            Tushar

  2. Ram
    May 8th, 2013 at 08:30 | #9

    Hi Tushar,

    I would like to verify the following, as i rebooted the server an started the Fusions Apps processes.

    I am unable to open Admin Console of Financial Domain
    http://fusionapps:18623/console

    I am unable to open other Domain urls as well.
    Am i missing something?

    I did the following.

    1. Start Listener
    lsnrctl start LISTENER_FUSIONDB

    2. Start Database
    sqlplus / as sysdba
    SQL> startup

    3. Start OID
    /app/fusion/admin/oid1/bin/opmnctl startall

    4. Start Web
    /app/fusion/admin/web1/bin/opmnctl startall

    5. Start Weblogic Admin Server for IDMDomain
    cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/
    nohup ./startWebLogic.sh &

    6. Start Weblogic Admin Server for IDMDomain
    cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/
    nohup ./startWebLogic.sh &

    7. Start Node Manager
    cd /app/fusion/fmw/wlserver_10.3/server/bin/
    nohup ./startNodeManager.sh &

    8. Start OAM managed serve
    cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/
    nohup ./startManagedWebLogic.sh wls_oam1 &

    9. Start remaining managed servers
    cd /app/fusion/admin/IDMDomain/aserver/IDMDomain/bin/
    nohup ./startManagedWebLogic.sh wls_ods1 &
    nohup ./startManagedWebLogic.sh wls_oim1 &
    nohup ./startManagedWebLogic.sh wls_soa1 &

    10. Start Weblogic Admin Server for other Domains

    nohup /app/fusion/instance/domains/fusionapps/CommonDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/FinancialDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/ProcurementDomain/bin/startWebLogic.sh
    nohup /app/fusion/instance/domains/fusionapps/BIDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/SCMDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/ProjectsDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/CRMDomain/bin/startWebLogic.sh &
    nohup /app/fusion/instance/domains/fusionapps/HCMDomain/bin/startWebLogic.sh &

    Regards,
    Ram

    • tushar
      May 8th, 2013 at 09:57 | #10

      Dear Ram,

      Make sure you startup all componets as per following post.
      http://www.oratraining.com/blog/2013/05/startupshutdown-scripts-for-fusion-applications-2-nodes-setup/

      Also when you start the weblogic admin servers, did you verify that that it is coming into RUNNING mode in the AdminServer.log file at /Servers/AdminServer/logs directory? You must wait for it to come to running state.

      Also make sure you are using correct ports as per summary after finishing the installation since your ports would be different than ports in my post. You can verify the same in /config/config.xml as well.

      Regards
      Tushar

      • Ram
        May 8th, 2013 at 21:50 | #11

        Hi Tushar,

        Thanks for info.

        I used the scripts as mentioned in
        http://www.oratraining.com/blog/2013/05/startupshutdown-scripts-for-fusion-applications-2-nodes-setup/

        When starting url http://fusionapps:18623/console (Financial Domain)
        i got Failure of server APACHE bridge:
        ——————————————————————————–
        No backend server available for connection: timed out after 10 seconds or idempotent set to OFF or method not idempotent.

        I didnot have this error before the reboot.

        Any suggestions?

        Regards,
        Ram

        • tushar
          May 9th, 2013 at 06:52 | #12

          Dear Ram,

          The fastartstop script starts the admin servers, start managed servers and then at the end stops admin servers, keeping managed servers up.

          So please check whether you are able to login to the homepage or not instead of admin console. If you want then you can manually start admin console using startWeblogic.sh scripts

          Thanks
          Tushar

  3. Daniel
    Apr 24th, 2013 at 12:27 | #13

    Hi Tushar,

    my postconfigure phase (SCM Installation) always fails during uploading the taxonomy (line 355 in fs-app-build.xml).

    I get ClassNotFoundException first:

    SEVERE: Error loading the JMX FRamework SPI “oracle.as.jmx.framework.standardmbeans.spi.JMXFrameworkProviderImpl”
    java.lang.ClassNotFoundException: oracle.as.jmx.framework.standardmbeans.spi.JMXFrameworkProviderImpl
    at java.net.URLClassLoader$1.run(URLClassLoader.java:202)….

    followed by:

    INFO: WSM-09004 Component auditing cannot be initialized.
    SEVERE: Error while invoking endpoint “http://vfusion01:15004/topologyManagerService/ProvisionConfigurationService” from client
    [SRC_CLASS: oracle.apps.fnd.provisioning.ant.taskdefs.util.logger.ProvisioningLogger] [SRC_METHOD: log] Client received SOAP Fault from server : PolicySet Invalid: WSM-02557 oracle.wsm.policymanager.accessor.BeanAccessor Query “&(@appliesTo~=”WS-Service()”)(policysets:global/%)” is queued for late retrieval. WSM-02557 oracle.wsm.policymanager.accessor.BeanAccessor Query “/policies/oracle/no_authentication_service_policy” is queued for late retrieval.

    ==
    It seems to be a policy issue, but I cannot figure out what could be wrong.

    Do you have any idea?

    Thanks & best regards
    Daniel

    • tushar
      Apr 24th, 2013 at 12:48 | #14

      Dear Daniel,

      Are you installing it on Linux or Solaris? Also is this 11.1.5 or 6? Also are you setting JAVA HOME before you run provisioning wizard? I remember having faced this on 11.1.1 on solaris and had spent huge amount of time manually getting this class jar file but the issue was something trivial. To be honest don’t recall right now but will go through my logs to find solution. Also make sure that you are manually able to launch complete url in browser and get xml output.

      Also if taxonomy upload fails, hcm and financial also will fail since they can proceed only if this step successfully finishes. Please do not skip any step by manually modifying build xml since it will make installation incomplete.

      Regards,
      Tushar

      • Daniel
        Apr 24th, 2013 at 13:28 | #15

        Dear Tushar,

        Thank you for reply.

        It is a Linux 11.1.6 one host installation. JAVA_HOME is defined.
        WSDL is accessible in the browser.

        I have got 96 GB RAM, which should be enough…

        Thanks & Best regards,
        Daniel

      • Daniel
        May 8th, 2013 at 06:49 | #16

        Hi Tushar,

        do you have any idea about this case?

        Thanks a lot & Best regards
        Daniel

  4. JL
    Apr 20th, 2013 at 07:35 | #17

    I used the fmwhost for IDM. When I got to PostConfigure, after it ran for 8 hours, everything failed. I did a retry and it complain about ldap on fmwhost. At that point, I tried logging into oam (http://fmwhost:7777/console), I could not log on. At that point, I realized that the postconfigure phase has hose the fmwhost. I could not log into oam without reboot, but it was only a matter of time that I had to repeat the whole process again.

    What do you suggest I should do?

    Thanks,

Leave a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>