Good CEST morning. I try to run two instances out of one oracle_home on a Windows 2008R2 server (bad OS, but not my choice). Problems are, that all databases are using the DIAG directory (crates a new subdir with the correct SID anyway) configured by the instance that has been configured last. And the second instance fails with ugly semaphore errors. I think they try to use the same ressource for the same reason they use the same DIAG dir. Fri Aug 27 08:30:30 2010 RECO started with pid=15, OS id=4896 Fri Aug 27 08:30:30 2010 MMON started with pid=16, OS id=5104 starting up 1 dispatcher(s) for network address '(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'... starting up 1 shared server(s) ... ORACLE_BASE from environment = D:\app\oracle Fri Aug 27 08:30:30 2010 alter database mount exclusive ORA-09341: scumnt: Datenbank kann nicht mit MOUNT angeschlossen werden OSD-04400: Interne Semaphore für Prozess kann nicht abgerufen werden O/S-Error: (OS 183) Eine Datei kann nicht erstellt werden, wenn sie bereits vorhanden ist. ORA-1102 signalled during: alter database mount exclusive... My questions are: I have no ORACLE_SID configured on system level. How does the box determine the oracle_sid used for a process in services.msc? Second: After establishing the instances with DBCA, everything was well, the insatnces did run simultanously. After stopping/starting both services, the mess described above happened. Do I have a wrong setup somehow? I want to use one oracle_home, because I want to avoid different patch levels and messing around with setting/changing ORACLE_HOME all the time. Thanks in advance! Martin -- Usn's IT Blog for Linux, Oracle, Asterisk http://www.usn-it.de -- //www.freelists.org/webpage/oracle-l