We shall be concentrating on the best pratices for managing Apps instances. If you come to know or would like to share the best practices you are most welcome to share.
Start/Stop Apps Instance.
This looks very simple, hah I know this is just run adstrtal.sh/adstpall.sh, addbctl.sh and addlnctl.sh. I used to think like this, then after gaining experience I came with my own steps.
In previous approach, Did I took adequate step for checking any running concurrent requests .. did I took enough precaution to check what critical happenings going on in database like hot backup.. Now we shall take care :-)
I assume you know the script location, I shall cover, if you still suggest me to include.
Starting is Simple.
- addbctl.sh start
- addlnctl.sh start SID
- adstrtal.sh apps/password
I start my preparation sometime before the downtime scheduled, to let the concurrent request finish. Following are the steps to bring down middle-tier services
- Bring down the concurrent manager before maintenance say 20 mins before.
adcmctl.sh stop apps/Password - Check if any concurrent reqeust is running. if running, check what it is doing, like sql, session is active.
- Check previous execution of similar program took how much time.
- If it is affecting downtime then login from front-end and terminate the concurrent program, and make a note of request id(communicate to user who submitted this request so they can submit again)
- Check the OS process id, whether it got terminated or not. If running then its a runaway process kill it. I dont like killings but...
SQL> select oracle_process_id from fnd_concurrent_requests where request_id=&Request_id;
- Check if hot backup is going on or not..
To check, go to alert log file $ORACLE_HOME/admin/CONTEXT_NAME/bdump/alert_sid.log
and also from sqlplus
SQL> select distinct status from v$backup;
If it returns row containing "ACTIVE" then hot back is in progress.
Wait till it gets over.
Otherwise next startup shall create problem. - Conditional - If you are using DR, pls take care of following steps
- Check which archive dest state refer for DR, enable it .
From show parameter log_archive_dest.. you may come to know..
say if you are using 3rd then run the sql
SQL>alter system set log_archive_dest_state_3=enable; - Check if standby is performing managed recovery.
SQL> SELECT PROCESS, STATUS FROM V$MANAGED_STANDBY;
PROCESS STATUS
------- ------------
ARCH CLOSING
ARCH CONNECTED
MRP0 WAIT_FOR_LOG
RFS WRITING
RFS RECEIVING
RFS RECEIVING - Cancel managed recovery operations.
SQL> ALTER DATABASE RECOVER MANAGED STANDBY DATABASE CANCEL; - Shut down the standby database.
SQL> SHUTDOWN IMMEDIATE;
- Check which archive dest state refer for DR, enable it .
- Stop listener
- Now stop the database.
- If still it is not going down, check in alert log , what exactly is going on.
- Check if any processes are running having local=n is running. If yes, kill..
