# Oracle EBS Concurrent Manager Not Starting: Troubleshooting Checklist > Source: https://netray.co/tools/oracle-ebs-concurrent-manager-not-starting-checklist > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Free Tools ## Short answer A step-by-step checklist for EBS 12.1 and 12.2 when the Internal Concurrent Manager or standard managers will not start. ## Category oracle ## Erp Oracle E-Business Suite ## Audience - Apps DBA - EBS system administrator ## Tool **Kind:** checklist **Groups:** #### 1. Confirm it is really down **Items:** Run adcmctl.sh status apps/ from $ADMIN_SCRIPTS_HOME and note whether the Internal Concurrent Manager is reported as running Open System Administrator > Concurrent > Manager > Administer and compare Actual vs Target processes for Internal Manager and Standard Manager Check OS processes on each concurrent node: ps -ef | grep FNDLIBR and ps -ef | grep FNDSM Confirm requests are actually stuck in Pending Normal or Pending Standby, not just slow #### 2. Read the logs **Items:** **Text:** Open the ICM log in $APPLCSF/$APPLLOG, named _MMDD.mgr, and read from the last startup attempt **Critical:** true Check the Service Manager log for the node (FNDSM log in the same directory) Check individual manager logs (w.mgr) and service component logs (FNDCPGSC.txt) Look for database errors (ORA-), node name mismatches and 'could not contact Service Manager' messages **Text:** Check that $APPLCSF/$APPLLOG and $APPLCSF/$APPLOUT exist, are writable and the file system is not full **Critical:** true #### 3. Database side **Items:** Query FND_CONCURRENT_QUEUES for CONCURRENT_QUEUE_NAME, CONTROL_CODE, RUNNING_PROCESSES, MAX_PROCESSES, TARGET_NODE Compare RUNNING_PROCESSES with MAX_PROCESSES; a nonzero running count with no OS processes indicates stale rows Query FND_CONCURRENT_PROCESSES for rows with PROCESS_STATUS_CODE still marked active for managers that are not running Check for locks on FND_CONCURRENT_QUEUES and FND_CONCURRENT_REQUESTS held by old sessions (V$LOCK / DBA_BLOCKERS) Confirm the database and its listener are up and the APPS password used by adcmctl.sh is current #### 4. Node configuration **Items:** **Text:** Check FND_NODES: NODE_NAME values must match the real hostnames, especially after a clone **Critical:** true **Text:** After a clone, run FND_CONC_CLONE.SETUP_CLEAN and then AutoConfig on every tier so FND_NODES is rebuilt **Critical:** true Confirm a Service Manager is defined for each concurrent processing node Check the profile option Concurrent:GSM Enabled is Y If Parallel Concurrent Processing is used, check primary and secondary nodes on each manager definition #### 5. Network and listeners **Items:** Check the applications listener with adalnctl.sh status; it serves FNDFS for viewing logs and output Confirm tnsnames.ora and any ifile under $TNS_ADMIN point to the right database host and service Test connectivity from the concurrent node with sqlplus apps@ Check hostname resolution in /etc/hosts and DNS for every node name in FND_NODES #### 6. Recovery **Items:** **Text:** Stop all managers with adcmctl.sh stop and confirm no FNDLIBR processes remain before any cleanup **Critical:** true **Text:** Back up FND_CONCURRENT_QUEUES, FND_CONCURRENT_PROCESSES and FND_CONCURRENT_REQUESTS before running any cleanup **Critical:** true Try Concurrent Manager Recovery in Oracle Applications Manager before manual scripts **Text:** Run cmclean.sql only as described in My Oracle Support Doc ID 134007.1, with all managers down **Critical:** true Start with adcmctl.sh start apps/ and watch the ICM log Confirm the Internal Manager is running first, then the Standard Manager reaches its target processes #### 7. After start **Items:** Check the Conflict Resolution Manager is running, or requests with incompatibilities stay Pending Check the Output Post Processor (FNDCPOPP) is up, or XML Publisher output fails Check Workflow service components (mailer, agent listeners) in Workflow Manager Submit Active Users or another simple request and confirm it completes Normal **Scoring Note:** Work top to bottom. Do not run cmclean.sql or any FND table cleanup with an open critical item in the Recovery group; backups and a confirmed full stop come first. ## Content **Intro:** This checklist walks an EBS 12.1 or 12.2 apps DBA through a concurrent manager that will not start, from confirming the state to reading the ICM log, checking FND tables and node definitions, and recovering safely. Critical items are the ones that can corrupt or lose request history if skipped. **Sections:** #### What the concurrent manager is **Body:** In Oracle E-Business Suite, concurrent processing runs batch programs (reports, interfaces, purges) in the background. The Internal Concurrent Manager (ICM) controls all other managers; it reads FND_CONCURRENT_QUEUES to know the target processes for each manager and asks the Service Manager (FNDSM) on each node to start them. The Standard Manager runs most requests, the Conflict Resolution Manager releases requests whose incompatibilities are clear, and service components such as the Output Post Processor and Workflow Mailer run under the same framework. If the ICM cannot start, nothing else does, which is why the ICM log is the first place to look. #### Where to look in the system **Body:** The control script is adcmctl.sh in $ADMIN_SCRIPTS_HOME; on 12.2 it is in the run edition. Logs are in $APPLCSF/$APPLLOG: the ICM log is named after the database SID and date, and each manager writes its own log. In the forms interface, Administer Concurrent Managers under the System Administrator responsibility shows actual and target processes and lets you verify, restart or deactivate managers. In the database, FND_CONCURRENT_QUEUES holds manager definitions and state, FND_CONCURRENT_PROCESSES holds one row per manager process, and FND_NODES lists the nodes known to the instance. **Bullets:** - FND_CONCURRENT_QUEUES: CONTROL_CODE, RUNNING_PROCESSES, MAX_PROCESSES, TARGET_NODE - FND_CONCURRENT_PROCESSES: PROCESS_STATUS_CODE, OS_PROCESS_ID, CONCURRENT_QUEUE_ID - FND_NODES: NODE_NAME, SUPPORT_CP #### The usual causes **Body:** After a clone, the most common cause is FND_NODES still holding the source hostnames, so the ICM tries to start managers on a node that does not exist; clean it with FND_CONC_CLONE.SETUP_CLEAN and rerun AutoConfig. After a crash or kill, stale rows in FND_CONCURRENT_PROCESSES and nonzero RUNNING_PROCESSES make the ICM think managers are already up; Oracle's cmclean.sql (Doc ID 134007.1) resets them, but only with every manager stopped. Other causes: a full log file system, a changed APPS password not used in the start command, wrong TNS entries, and a stopped database listener. #### Common mistakes **Body:** Running cmclean.sql while some FNDLIBR processes are still alive leaves orphan processes that keep processing requests outside the ICM's control. Deleting rows from FND_CONCURRENT_REQUESTS to clear a backlog destroys audit history; cancel requests instead. Restarting the whole application tier when only the concurrent tier is affected wastes time. Finally, do not change MAX_PROCESSES directly in the table; use Define Concurrent Manager work shifts so the change is recorded. **Cta:** If you want your support team to ask EBS what is stuck in concurrent processing in plain English, on your own servers and read-only, we can set that up. **Faq:** #### What is the concurrent manager in Oracle EBS? **Answer:** It is the framework that runs background programs. The Internal Concurrent Manager controls the others, Service Managers start processes on each node, the Standard Manager runs most requests and the Conflict Resolution Manager handles incompatibilities. Requests are stored in FND_CONCURRENT_REQUESTS and queued to managers by specialization rules. #### How do I check concurrent manager status from the command line? **Answer:** Run adcmctl.sh status apps/ from $ADMIN_SCRIPTS_HOME. It reports whether the Internal Concurrent Manager is running. Combine it with ps -ef | grep FNDLIBR on each node and the Administer Concurrent Managers form to see actual and target processes. #### Is it safe to run cmclean.sql? **Answer:** It is Oracle's non-destructive script for resetting concurrent manager tables, documented in Doc ID 134007.1, but only when all managers are stopped and no FNDLIBR processes remain. Back up the FND concurrent tables first and use the version from the current note. #### How do I start the applications listener in EBS 12.2? **Answer:** Use adalnctl.sh start from $ADMIN_SCRIPTS_HOME of the run edition, or adstrtal.sh to start all application tier services. The applications listener serves FNDFS, which you need to view concurrent request logs and output from the browser. ## Search terms oracle ebs concurrent manager not starting, what is concurrent manager in oracle ebs, oracle ebs 12.2.12 start listener, oracle ebs concurrent program status code, adcmctl.sh status, internal concurrent manager not starting, cmclean.sql ebs --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json