We have successfully delivered one database application to one customer, on Oracle 10gR2 server (v10.2.0.2, Unix, 64-bit).
Since this is a RAC environment and because of some problems within one stored procedure we have implemented ("ORA-01801: date format too long for internal buffer"), which happens from time to time only in the last couple of weeks, we have asked to get access rights for two admin views, V$INSTANCE and GV$INSTANCE.
My idea was that this problem can be caused by one Oracle instance (host machine) which is not properly configured and/or has a malfunction.
After checking the view GV$INSTANCE, there is only one row returned with inst_id = 1.
What does that actually mean? Is this then really a RAC environment or only a single Oracle instance? Or it is possible to limit the result of this view, in order to prevent the user to check/examine the system configuration?
Thanks in advance.
1. The instance is configured to be non-RAC
2. the parallel_max_servers parameter is set to 0.
You can check the PARALLEL value in the v$instance views on both nodes to determine whether the instance thinks it is part of a cluster.
You can check the parallel_max_servers parameter via the v$parameter view or by doing a show parameter command from SQL*Plus.
My guess is that the ORA-01801 is not related to RAC. If this is your only customer receving the error, I would suspect bad data first, an Oracle bug second.
Also, it is not possible to "fool" the user via the gv$instance view without hacking the data dictionary or creating a view or synonym in the users schema. I doubt this is happening.
I would advise focusing on the data manipulated in the procedure.