Oracle EBS-Blank Page Appslogin/AppsLocal Login Page in R12

There were quite a few times that I have encountered this blank front-end issue. The root cause was not the same every time it was investigated. There are many reasons for this issue to occur which I have compiled to make it easier for Apps DBA’s. Listed below are some of them.

Here is the error page while logging into Oracle E-Business Suite front end.

These are some of the main reasons for the APPS blank login page in R12:

·    JSP pages not compiled properly after the clone

·    Data block corruption on Cloned environment with bulk synchronization issues

·    Database Listeners are down

·    Archive log space is full

·    DB or APPS file system is Full including /tmp location

·    Java Cache issues between the multiple middle-tiers

 

Here are different scenarios and work around/Fixes:

Scenario 1:  After the clone if any of the users experience user login issues other than super user with the below error message then please follow the steps in scenario 1.

The login in page may show Http-500 Internal server error (or) Http-404

1.         http 500 internal server error:

              java.lang.NoClassDefFoundError
              at oracle.apps.fnd.sso.AppsLoginRedirect.doGet(AppsLoginRedirect.java:623)
              at javax.servlet.http.HttpServlet.service(HttpServlet.java:743)
              at javax.servlet.http.HttpServlet.service(HttpServlet.java:856)
              at com.evermind[Oracle Containers for J2EE 10g (10.1.3.0.0)].

              server.http.ResourceFilterChain.doFilter(ResourceFilterChain.java:64)
              at oracle.apps.jtf.base.session.ReleaseResFilter.doFilter(ReleaseResFilter.java:26)

Solution: (from Application Tier)

$cd $ADMIN_SCRIPTS_HOME $adopmnctl.sh stopall — to shutdown opmn managed services viz., opmn, forms, oacore & oc4j  $ cd $COMMON_TOP/_pages $ mv _pages _pages_old $mkdir _pages $ cd $FND_TOP/patch/115/bin or export PATH=$FND_TOP/patch/115/bin:$PATH Compile all jsps. $perl ojspCompile.pl –compile –flush -p 2 (or) ojspCompile.pl --compile --fast --quiet (-p represents parallel threads, if you have enough horse power on your server, you can increase it to 8 or 10) $cd $ADMIN_SCRIPTS_HOME adopmnctl.sh startall

OR:

ORA-01578: ORACLE data block corrupted (file # 23, block # 221085)
ORA-01110: data file 23: ‘/RDBMS ORACLE_HOME/app/a_txn_data02.dbf’
ORA-26040: Data block was loaded using the NOLOGGING option

Solution: Login as apps user

$ sqlplus apps/apps_pwd

set serveroutput on size 100000;  declare  begin  WF_LOCAL_SYNCH.BULKSYNCHRONIZATION(  P_ORIG_SYSTEM => 'ALL',  P_PARALLEL_PROCESSES => null,  P_LOGGING => null,  P_RAISEERRORS => TRUE);  exception  when others then  dbms_output.put_line('sqlerrm = ' || sqlerrm);  end;  /

Scenario 2: Database Listener is down

Solution:  Start the database listener.

Scenario 3: Database Archive Locaion is 100% full or any mount point or tmp location is full(including concurrent log and output storage mount point)

Solution: Either temporarily move the files/archivelogs to another location or delete older files/archives (provided you had taken a cold backup sometime).

Scenario 4: IAS Cache initialization failed (Doc ID 1484840.1)

E-Business Suite Applications Technology Stack,  HTTP / IAS Server related issues

Users report cache initialization errors in OA Framework related pages.

The oacorestd.err file shows: 

Exception in static block of jtf.cache.CacheManager. Stack trace is: oracle.apps.jtf.base.resources.FrameworkException: IAS Cache initialization failed. 

The Distributed Caching System failed to initialize on port: 12360. The list of hosts in the distributed caching system is: 10.24.248.41 . 

The port 12360 should be free on each host running the JVMs. 

Solution:

To resolve the issue test the following steps in a development instance and then migrate accordingly:

I. For Autoconfig enabled environment:
   1. Bring down opmn services on all boxes.
   2. Verify no process is listening for s_java_object_cache_port value:

  netstat -an | grep [port number] | grep LISTEN

    Identify the OS process associated with a port:

  lsof -i tcp:[port number]

   3. Change s_java_object_cache_port on all the nodes in $CONTEXT_FILE.

   4. Run autoconfig.

II. For Non-Autoconfig enabled environments:

  s_java_object_cache_port value is referrenced in two places

   1. $INST_TOP/ora/10.1.3/javacache/admin/javacache.xml

   2.  JTF_DIST_CACHE_PORT profile option.

   Change the port value to a new number in the $CONTEXT_FILE and bounce opmnctl services.

   3. Verify no process is listening for s_java_object_cache_port value : 

  netstat -an | grep [port number] | grep LISTEN

   Identify the OS process associated with a port

  lsof -i tcp:[port number]

If the newly assigned port still does not listen on the server you can use workaround given bellow. This workaround will resolve the “IAS Cache initialization failed” even if the port is not listening.

 Workaround: (Note: The below workaround is intended for testing purposes only to help identify RCA)

a. Set LONG_RUNNING_JVM=false in $INST_TOP/ora/10.1.3/j2ee/oacore/config/oc4j.properties.

b. Bounce opmn services.
Note: This happens when Distributed Java Caching is enabled in a multi middle tier environment

Note: In 11i, if the database listener is down or if the archive location is full, may get the internal server page while trying to access the login page. In R12 may get a blank page and apache error log will show the “oc4j_socket_recvfull timed out” and "mod_oc4j: request to OC4J [mid-tier servername:ajp port] failed: Connect failed" error messages.

Scenario 5: mod_oc4j: Failed to find a failover oc4j process for session request for destination

Error Noticed in the log file 

$LOG_HOME/ora/10.1.3/Apache

Solution: 

1. Noticed the When we have started the Application Services the load average was 1.

2. When executed ps -ef  |grep applmgr  noticed many processes.

3. Killed the process via 

ps -ef |grep applmgr|grep -v grep |awk '{print $2}'|xargs kill -9

Issue got resolved.

a. Started the services.

b. Via forms fixed restart and all the concurrent managers are up.