Simple Concurrent Manger Issue

Below are few steps to restart concurrent manager; in order to trouble shoot simple issue in a unix based env.

1. login to form
http://:/dev60cgi/f60cgi
as sysadmin

2. select system administrator responsibility

3. concurrent -> manager -> Administer
will pop up the window with the status of the manager.

4. check for the standard manager
whether actual = target
1) if target=0, thats mean standard manager is down.
2) if target=some number, thats mean standard manager is up

whether actual <> target

if actual = 0 and target > 0

standard manager has not come up properly.
for cross verifying it. submit a reuqest for "active user". it should get completed within a min or two. wait for few min and refresh the request. if its still penfding or showning inactive manager, check the concurrent -> manager -> Administer window against the standard manager , in pending request active user req will be there.

now open an sqlplus session and execute the script ccm.sql and cm_stat.sql

statistics can be gathered for the concurrent process and FND tables.

5. now login to unix prompt as application user( a like a) and move to directory $SCRIPTS_TOP

execute adcmctl.sh stop apps/passwd

check for the concurrent processes
ps -ef grep FND grep
should continously go down.

6. meanwhile keep on checking concurrent -> manager -> Administer window
All the entry target/actual should go down to 0. pending can be there with the exception for the internal manager.

7. wait for some time. (20 min)
if its not coming down to 0 even waiting for 5 mins or so. First cancel the request for the shutdown.
and kill the process on unix box.

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

8. from the sqlplus session cm_stat.sql , should get 0 actual and 0 target for all the managers.

also from concurrent -> manager -> Administer window
All the entries should be 0. Pending can be there with the exception for the internal manager.

9. Now from sqlplus session:
Run cmclean.sql
it'll ask for typing dual. Type dual and it'll start updating the corrupted FND tables entries.
Once the script is through type commit

10. and restart the concurrent manager from the unix prompt.
adcmctl.sh start apps/passwd
check whether the no of FND process is going up by command below:

ps -ef grep FND grep RPAD11

Also from concurrent -> manager -> Administer window refresh the page.

This time no of standard manager should come up to the number mentioned in target.
if not wait for another few min and refresh the administer page.
issue should get resolved.

Note:
Concurrent Manager Log and Out files are generated at following location:

LOG file: $APPLCSF/$APPLLOG
OUT file: $APPLCSF/$APPLOUT

e.g. typical locations may be:
/u01//sqlcom/log/_
/u01//sqlcom/out/_