R12.1.1 / 11.1.0.7 upgrade : Error with 10.1.3.4 patchset upgrade

Question

All,
We are on 12.0.2 / 10.2.0.2 (non-rac) in a split-tier architecture. Web/Forms on the Linux node and Concurrent Managers/Database on the Solaris node. We are now on a project to upgrade this environment to 12.1.1 on 11.1.0.7 db. (Non RAC,Non ASM)
Here are the high level tasks that i laid down
-- Create a project instance (clone from Prod)
-- Upgrade db to 10.2.0.4 from 10.2.0.2
-- Upgrade the techstack Oracle Homes (10.1.2 and 10.1.3)
-- Upgrade the JDK to 1.6.0_X
-- Upgrade 12.0.2 to R12.1.1
-- Upgrade the database to 11.1.0.7
PS: Direct ugprade from 10.2.0.2 to 11.1.0.7 needs us to be on R12 RUP4. So, i decided to do 10g upgrade as an intermediate step. We are following 752619.1 and all the notes that it refers.
We are hitting an issue upgrading the 10.1.3 Oracle home on the Solaris node. The same upgrade completed fine on the Linux node. Here is the description of the issue.
The runInstaller for 10.1.3.4 patchset expects OPMN to be running for a successful patchset installation. On a node with just batch processing enabled, OPMN does not run by default. The installer tries to start OPMN and errors out saying
========
OPMN is failing to start. OPMN must be started in order to correctly install and configure the patchset. Please start OPMN ma
nually in order to continue with the patchset installation.
=========
This is when the installer is at 83%
We tried to start OPMN and managed processes manually using adopmnctl.sh
OPMN starts and is able to start the HTTP server too but not any of the OC4J container components (oacore,forms and oafm). Here are the errors
==========
ias-instance id=afcprj_swafcdev1.swafcdev1.sherwin.com
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
ias-component/process-type/process-set:
default_group/oacore/default_group/
Error
--> Process (index=1,uid=1861636864,pid=17677)
failed to start a managed process after the maximum retry limit
Log:
/swafc/applmgr/afcprj/inst/apps/afcprj_swafcdev1/logs/ora/10.1.3/opmn/default_gr
oup~oacore~default_group~1
ias-component/process-type/process-set:
default_group/forms/default_group/
=========
The logfile that it points to does not have any errors. Looking into the other log files within $LOG_HOME/ora/*3/
, we get the following error in oacorestd.err
=========
09/10/05 16:07:10 Error initializing server: Shared library "oracle.ws.jaxrpc" could not be found.
09/10/05 16:07:15 Error: <connector name="OracleASjms" path="OracleASjms.rar" /> will not be bootstrapped since corresponding module declaration was not found in application.xml.
09/10/05 16:07:17 Error initializing server: Shared library "oracle.ws.jaxrpc" could not be found.
========
I saw some notes and tried updating the config xml files and removed the references to oracle.ws.jaxrpc but to no help.
Installer log has the below error when we attempt the "retry" operation in installation screens.
========
fileName = NonExistentFile
searchString = NonExistentString
Query Exception: FileNotFoundException
Query Exception Class: class oracle.sysman.oii.oiil.OiilQueryException
Query Exception: FileNotFoundException
Query Exception Class: class oracle.sysman.oii.oiil.OiilQueryException
===========
Any help on the above issue? We are never going to use OPMN on this node but stuck here currently !!
Deviating a little bit, if we have to consider moving the concurrent managers out of database node to the
web/forms node (just to avoid having to maintain 2 appl_tops and getting away with doing the above tasks 2 times)
, what are the various things to consider?
-- Network latency
-- R12 being Unified APPL_TOP, can we just try to bring up ICM on Web/Forms node? (just enable the service)
-- We are aware that we might have to move the custom top and test custom programs (especially host programs).
Is there any other considerations when moving Concurrent managers from a Solaris node(64 bit) to Linux node (32 bit) ?
One last question: Why are these adoacore/adoafm scripts taking a long time?
Any feedback is highly appreciated.
Thanks,
Arun

Answer 1

Hi,
Error
--> Process (index=1,uid=1861636864,pid=17677)
failed to start a managed process after the maximum retry limit
Log:
/swafc/applmgr/afcprj/inst/apps/afcprj_swafcdev1/logs/ora/10.1.3/opmn/default_gr
oup~oacore~default_group~1
ias-component/process-type/process-set:
default_group/forms/default_group/Please try the suggested solution in Note: 555214.1 and see if it helps.
Query Exception: FileNotFoundException
Query Exception Class: class oracle.sysman.oii.oiil.OiilQueryExceptionCould you please verify that permissions are set correctly on the application tier node files?
-- R12 being Unified APPL_TOP, can we just try to bring up ICM on Web/Forms node? (just enable the service)Please see (Note: 373611.1 - How to move Concurrent Processing Server from one node to another node) -- Step 3 to 7
-- We are aware that we might have to move the custom top and test custom programs (especially host programs).
Is there any other considerations when moving Concurrent managers from a Solaris node(64 bit) to Linux node (32 bit) ?I would suggest you relink all executable files via adadmin and check then.
Regards,
Hussein