RMAN suggestions sought - Sybrand, Thomas Kyte, Anyone?

    I've been learning about RMAN with an NT server running a database (the target) and an NT workstation running a small catalog database and OEM. In line mode everything is working. And when I run through OEM in immediate mode everything works as well.

    But running through OEM by scheduling RMAN jobs is not working. If I schedule a simple job such as a target/catalog resync (6 seconds in line mode) to fire every ten minutes through OEM, sometimes the job fires not at all, sometimes once then not again, sometimes 100 times then stops. Oracle support has been doing all it can for three days, with no luck so far. Do any of you have any suggestions, ideas or thoughts while Oracle pores over the trace files I sent?

Thanks,

Van

PS We've already blown away the recovery catalog and recreated it, got rid of all intelligent agent files and re-created them, rediscovered the nodes and made sure preferred credentials were established for the nodes and for the databases, checked all tnsnames.ora listener.ora files, tried it with a second NT server with the same results, made sure machine names and machine names with domains were in the etc/hosts files, made sure the machine names could be resolved by the DNS server, made sure the workstation and server were synchronized to each other through an atomic clock program, made sure the user(s) were given enough privileges in NT user manager including "logon as batch job", stopped and restarted the databases. Received on Sun Apr 02 2000 - 18:31:13 CDT