Concurrent Manager (CM) Overview and Real-Time Handling – Oracle EBS
Overview of Concurrent Manager (CM):
Concurrent Manager is an important part of Oracle E-Business Suite. It is responsible for running background jobs and tasks submitted by users. These tasks are called concurrent requests and can be reports, PL/SQL programs, or system-related jobs. CM makes sure these jobs are handled properly without overloading the system. There are different types of managers like Standard Manager, Inventory CM, and Internal Manager, each handling different kinds of jobs.How a User Request Becomes a CM Job:
When a user submits a job from the front-end (like Payroll or a report), Oracle creates a concurrent request for it. This gets stored in a table (FND_CONCURRENT_REQUESTS). The request includes details like program name, parameters, and status. The Internal Manager assigns the job to the right CM based on setup. Some programs also trigger multiple sub-requests or child requests. The actual executable linked to the request runs in the background.How CM Handles the Request:
The CM picks up the request from the queue, starts a background process, and
runs the program. Each job runs as a separate OS process in the background.
Some programs are simple, but others like Payroll might involve multiple steps
and sub-processes. These can be tracked in the database by checking the parent
and child request IDs. You can find these in FND_CONCURRENT_REQUESTS and related tables.
Job Monitoring and CM Availability:
You can monitor job progress from the application or by querying the DB directly. If CM is down, new jobs won’t start. Running jobs may fail or get stuck. It's important to check for any pending or failed jobs after a CM downtime. Jobs with statuses like “Error” or “Terminated” should be checked carefully before rerunning.Why CM May Go Down and Link to WF Mailer:
There are many reasons for CM to stop working: disk full, database issues, stuck jobs, or resource problems. Log files under $APPLCSF/log (e.g., FNDCPGSC*.log) help find the issue. While CM and Workflow Mailer are separate services, they both rely on the same database and environment. If CM is down, any workflow-related programs won’t run. If the WF Mailer is down, notifications won’t be sent, but CMs will still run jobs.Restarting CM and What Happens to
Jobs:
If CM is down due to an issue (like space full), we fix the problem and start
CM again using the script:
$ADMIN_SCRIPTS_HOME/adcmctl.sh
start
CM will pick up new and scheduled
requests. But jobs that failed during the crash won’t resume automatically.
They might show as failed or terminated. We need to check logs for each failed
request, understand the reason, and then decide whether to rerun. We should
avoid blindly rerunning jobs.
Post-CM Restart Actions:
Once CM is restarted, we should:- Check for failed requests in FND_CONCURRENT_REQUESTS
- Open and review log files for those jobs
- Coordinate with the functional team if necessary
- Re-submit jobs only after confirmation and analysis
This ensures we don’t create
duplicate data or cause business issues. Every job behaves differently, so
reviewing each one after CM is restarted is very important.
Comments
Post a Comment