Issue with Concurrent Manager

Be not afraid of growing slowly, be afraid only of standing still.

Chinese Proverb

CM– Concurrent Manager

Most of the DBA face the problem with CM.
CM is heart of Oracle APPS

Procedure to start CM:

1. Start the Database.
2. Start the Apps Tier

• Unix user – apstartal.sh apps/apps
• Windows user – use SCM to start the service or adstrtal.cmd apps/apps
• Script is located under $COMMON_TOP/admin/scripts for 11i and $INST_TOP/admin/scripts for R12
• Log file location : $APPLCSF/$APPLLOG

Steps to debug the CM :

1. Check the log file CM . CM will not start if unable to write the information in log file.
2. Tnsping $TWO_TASK . Two task is nothing but a sid name . Check its executing successfully.
3. Check for the process FNDLIBR. Unclean shutdown will create a problem.
4. Check for the FND invalids in database
5. Keep an eye on alert logs of database . It will cause an core dump issue

How the CM works :

1. First it will start the ICM(Internal concurrent manager) and ICM will start remaining managers like standard managers..etc
2. FNDSM is a process is responsible for starting other managers.

Features involved in CM:

1. You can create any number of managers and assign a workshift to the managers.
2. CM can handle any number of request depends on the server configuration.
3. If CM wants to handle more process(FNDLIBR). Increase the count of FNDLIBR process as a system administrator and bounce the CM

Real time example:

1. CM will connect to database by resolving tnsnames.ora . After changing the ip of the server(multi node installation) i cannot able to up the CM . Problem because /etc/hosts entry still pointing to the old ip after changing the /etc/host file CM is up successfully .It sounds simple but tricky to debug . Finally tnsping utility helped me to analyse the problem

Related Terms:
Fndsm,fndlibr,concsub,palibr,adcmctl,cpmgr