Tuesday, March 31, 2009

R12 EBS Installation Tips and Tricks: Failed Installation Solutions for RW-50015 Errors

After countless hours of frustrations with attempts to get a test Oracle EBS R12 installed on my Macbook Pro laptop running VMWare Fusion and Linux, I found a solution recently to a problem with installing a test Oracle R12 EBS environment on Oracle Enterprise Linux 5 (OEL 5- 32 bit) with VMWare Fusion. There is a bug in R12 EBS for Linux platforms that will cause the installation to fail on the Apache HTTP server process.

What happens is that you will receive an RW-50015 Error message during the Rapidwiz install process.


RW-50015: Error: - HTTP Listener is not responding. The service might not have started on the port yet. Please check the service and use the retry button.


Also at same time, the Rapid Install window will show many errors for HTTP.


I then used adapcctl.sh status to check on these services and sure enough the http server would not start:

[oracle@ebs scripts]$ ./adapcctl.sh status

You are running adapcctl.sh version 120.6.12000000.2

Checking status of OPMN managed Oracle HTTP Server (OHS) instance ...

Processes in Instance: PROD_ebs.ebs.sandiego.com
-------------------+--------------------+---------+---------
ias-component | process-type | pid | status
-------------------+--------------------+---------+---------
OC4J | oafm | 21168 | Alive
OC4J | forms | 21078 | Alive
OC4J | oacore | 20943 | Alive
HTTP_Server | HTTP_Server | N/A | Down


adapcctl.sh: exiting with status 0

adapcctl.sh: check the logfile /d01/app/oracle/inst/apps/PROD_ebs/logs/appl/admin/log/adapcctl.txt for more information ...



According to the release notes from Oracle Metalink Support (Note #402310.1)
Oracle Applications Installation and Upgrade Notes Release 12 (12.0.4) for Linux (32-bit)
Oracle HTTP Server Patch (on OEL 5 and RHEL 5 only)

There is an issue with the Oracle HTTP Server that comes with the R12 EBS Suite that causes the linking process to fail for Apache. The solution is to exit the Rapid Install program, shutdown the OC4J and web services and apply the patch. Actually, it is really a library called libdb.so.2 which is missing from the default OEL 5 and RHEL 5 operating system. Even after I ran the up2date oracle-validated command to install all of the fixes for Oracle on Linux, it was still missing. So the patch really is to download the patch which contains this essential missing library file and copy it over to the /usr/lib directory structure.


Download and apply the patch 6078836 from OracleMetaLink to fix an issue with the Oracle HTTP Server bundled with the E-Business Suite technology stack. We unzip the patch:

[oracle@ebs oracle]$ unzip p6078836_101330_LINUX.zip
Archive: p6078836_101330_LINUX.zip
creating: 6078836/
inflating: 6078836/libdb.so.2
inflating: 6078836/README.txt
[oracle@ebs oracle]$ ls
6078836 apps db inst p6078836_101330_LINUX.zip StageR12
[oracle@ebs oracle]$ cd 6078836/

[oracle@ebs 6078836]$ ls
libdb.so.2 README.txt


Then we copy to /usr/lib.

[root@ebs 6078836] $ cp libdb.so.2 /usr/lib

[root@ebs lib]# cd /usr/lib

[root@ebs lib]# pwd
/usr/lib


[root@ebs lib]# ls libdb.so*
libdb.so libdb.so.2



After we apply the patch, we need to restart the web services using the adapcctl.sh utility. Actually, just re-run the Rapid Installer program which will do this step for us. After it is successful and Apache comes up, we can verify status for R12 EBS by running the adapcctl.sh command again as listed in the below example:

[oracle@ebs scripts]$ ./adapcctl.sh status

You are running adapcctl.sh version 120.6.12000000.2

Checking status of OPMN managed Oracle HTTP Server (OHS) instance ...

Processes in Instance: PROD_ebs.ebs.sandiego.com
-------------------+--------------------+---------+---------
ias-component | process-type | pid | status
-------------------+--------------------+---------+---------
OC4J | oafm | 21168 | Alive
OC4J | forms | 21078 | Alive
OC4J | oacore | 20943 | Alive
HTTP_Server | HTTP_Server | 28519 | Alive

We then re-start rapidwiz and voila! It works!




Now, we can access the R12 EBS Webportal main logon page as shown below:





We can now login to R12 EBS as SYSADMIN user using default sysadmin password as shown below to verify that all is working for our new R12 EBS installation:





Cheers,
Ben

5 comments:

Natarajan said...

Hello Ben where is adapcctl.sh located ?

Natarajan said...

Hey Ben how to shut down OC4J and web services ?

Natarajan said...

Hi Ben,
I did as per your instructions , and copied the libdb.so2 to /usr/lib but still when I restart all services my HTTP server exits out with 204. IS there something I should run along with copying that file ?

this is the log file i get
06/03/09-01:08:28 :: adapcctl.sh: starting OPMN managed OHS instance

opmnctl: starting opmn managed processes...
===================================
opmn id=lnx03.dreamzhut.com:6200
0 of 1 processes started.

ias-instance id=VIS_lnx03.lnx03.dreamzhut.com
+++++++++++++++++++++++++++++++++++-----------------------------------ias-component/process-type/process-set:
HTTP_Server/HTTP_Server/HTTP_Server/

Error
--> Process (index=1,uid=1283072333,pid=10025)
failed to start a managed process after the maximum retry limit
Log:
/d01/oracle/VIS/inst/apps/VIS_lnx03/logs/ora/10.1.3/opmn/HTTP_Server~1.log


06/03/09-01:08:32 :: adapcctl.sh: exiting with status 204

=================================

Please help me out
Thanks !

Ben Prusinski said...

Natarajan,

1) Look under $INST_TOP/ora/10.1.3/$APPLFENV or $ADMIN_SCRIPTS_HOME directory for adapcctl.sh

2) opmnctl stopall

3) did you follow instructions per the patch from Metalink? You need to stop the apps tier, apply the patch which is the library file, then restart the apps tier and it should work.

Check out Metalink Note 743518.1 for more details.

Cheers,
Ben

Hai Yan (Henry) Liang said...

thanks for your article.