How to get status of a backup job

Question

Hello everybody,
I scheduled our daily RMAN backup using the DBMS_SCHEDULER package. The DB job calls an external batch file which in turns calls rman's executable and passes it a command file containing the actual backup script. Similar to what's described here:
Problems with automated RMAN Backup in Oracle 11GR1 with DBMS_SCHEDULER
The backup works perfectly. The problem is that if I intentially corrupt the command file (simulating a backup failure) so that the backup fails, and then query the dba_scheduler_job_log view, I always get a status of SUCCEEDED. Regardless of the actual backup outcome, the job seems to always succeed. That's a problem for me, because I'd like to set up a user defined metric which would check the status of the job in dba_scheduler_job_log. If the job wasn't successfully executed, a mail would be sent.
Does anybody know why the job doesn't fail when it should? I also tried to execute the failing script manually and then checked the %ERRORLEVEL% environment variable. It has a value of 1, meaning error.
Btw, I know that if I schedule the same backup with Enterprise Manager's job system (instead of DBMS_SCHEDULER), the backup failure is correctly reported (and the metric works). The issue with EM is its limited scheduling capability - I do a full backup on Sunday, a cummulative incremental backup on Wednesday and all the other days a differential incremental backup. Since now I haven't found a way to define a schedule with an appropriate repeating interval. The only solution I figured out is to have 7 jobs, one for each day of the week....
Thanks in advance for any answers,
Jure

Answer 1

Hi,
@alanm: I completely forgot checking that view :-) Using it is a solution, but I should probably set up a job chain to handle that (run backup & chack status in the view afterwards)....if there isn't be another solution, maye I'll do that. But then, maybe using 7 EM jobs is a simpler solution.
@RnR: I tried to execute the batch file manually from the command line like that:
C:\>echo %errorlevel%
0
C:\>rman target / cmdfile=backup_full.rcv
Recovery Manager: Release 10.2.0.3.0 - Production on Sre Jul 2 09:46:37 2008
Copyright (c) 1982, 2005, Oracle. All rights reserved.
connected to target database: TESTDB (DBID=2890535676)
RMAN> RUN
2> {
3> aa;
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-00558: error encountered while parsing input commands
RMAN-01009: syntax error: found "identifier": expecting one of: "allocate, alter, backup, beginline,
blockrecover, catalog, change, copy, convert, crosscheck, configure, duplicate, debug, delete, exec
ute, endinline, flashback, host, mount, open, plsql, recover, release, replicate, report, restore, r
esync, set, setlimit, sql, switch, startup, shutdown, send, show, transport, validate"
RMAN-01008: the bad identifier was: aa
RMAN-01007: at line 3 column 1 file: backup_full.rcv
C:\>echo %errorlevel%
1
So the if the script fails, I get an errorlevel of 1. Is it possible that the status isn't propagated back to the Oracle job?
@alanm: I'm already appending to the log, but I don't see how can that affect the outcome status of the job?
Thanks for your answers.
Regards,
Jure