EBS R12.2 : Managed servers failed to start after server crash

By Gowthami | apps-dba.com | E-Business Suite Series

After an unexpected server crash or hard reboot, Oracle E-Business Suite R12.2 managed servers (WebLogic AdminServer, oacore, forms, oafm) often fail to start due to stale lock files, incomplete shutdowns, or corrupted state files. This post covers the diagnostic steps and fixes to get EBS back online quickly.

What You Will Learn: How to diagnose and resolve managed server startup failures in Oracle EBS R12.2 after a server crash, including clearing lock files, cleaning WebLogic state, and restarting services in the correct order.

Common Symptoms

• adstrtal.sh completes but managed servers show FAILED or UNKNOWN state
• WebLogic AdminServer starts but oacore/forms/oafm fail to start
• Error: RUNNING mode failed - server is already started
• Error: Lock file exists or PID file found

Step 1: Check Running Processes

First, verify that no stale Oracle/WebLogic processes are still running:

# Check for any running WLS or EBS processes
$ ps -ef | grep -E "weblogic|oacore|opmn|java" | grep -v grep

# Kill any stale processes found (replace PID)
$ kill -9 <PID>

Step 2: Remove Lock Files and PID Files

After a crash, lock files may prevent services from starting:

# Navigate to instance home
$ cd $INST_TOP

# Remove WLS lock files
$ find . -name "*.lck" -delete
$ find . -name "*.pid" -delete

# Remove OPMN lock files
$ rm -f $INST_TOP/ora/10.1.3/opmn/logs/*.pid
$ rm -f $INST_TOP/ora/10.1.3/opmn/logs/opmn.pid

Step 3: Clean WebLogic Server State

Corrupted WebLogic state files must be removed before restarting:

# Navigate to domain directory
$ cd $EBS_DOMAIN_HOME

# Remove server state directories (replace SERVER_NAME)
$ rm -rf servers/AdminServer/tmp
$ rm -rf servers/oacore_server1/tmp
$ rm -rf servers/forms_server1/tmp
$ rm -rf servers/oafm_server1/tmp

# Remove .lok files
$ find $EBS_DOMAIN_HOME -name "*.lok" -delete

Step 4: Clean Deadlock Files

$ rm -f $EBS_DOMAIN_HOME/servers/AdminServer/data/ldap/ldapfiles/EmbeddedLDAP.lok

Step 5: Restart in Correct Order

Always start EBS R12.2 services in this sequence:

1. Database (if not running): startup from SQL*Plus
2. Database listener: lsnrctl start
3. Application tier—run as applmgr:

$ source /u01/install/APPS/EBSapps.env run
$ $ADMIN_SCRIPTS_HOME/adstrtal.sh apps/<APPS_PASSWORD>

Step 6: Verify Managed Server Status

# Check status of all services
$ $ADMIN_SCRIPTS_HOME/adstatus.sh

# Or check via WebLogic console
# http://<hostname>:7001/console

Troubleshooting Tips

Issue	Likely Cause	Fix
AdminServer won't start	Stale .lok or .pid file	Remove lock files, clean tmp
oacore FAILED state	Corrupted server state	Remove servers/oacore*/tmp
OPMN not starting	Stale opmn.pid file	Delete opmn.pid and restart
Forms server errors	Database not accessible	Verify DB and listener are up first

Prevention

• Always use adstpall.sh for graceful shutdown before patching or rebooting
• Configure EBS services as OS startup scripts (rc.local or systemd) for automatic, ordered startup
• Monitor disk space—full filesystems are a leading cause of crash-related startup failures

Master Oracle Exadata

This post is part of our E-Business Suite administration series. Get our comprehensive Exadata guide with architecture, performance tuning, and practical DBA troubleshooting techniques.

Get the Exadata PDF Guide