Hmm, I don't see a FAILOVER = yes entry in your GWIT entry -- I have a 10g 2 node RAC cluster also. My test works for me -- that is sess1 sqlplus select, and shutdown abort the instance from sess 2. Sess 1 fails over to the surviving instance and finishes the select statement from where it left off...
Did you happen to verify how you were connected before perfoming the shutdown abort?
I used a slightly modded Jeff Hunter script to do this:
COLUMN instance_name FORMAT a13 COLUMN host_name FORMAT a9 COLUMN failover_method FORMAT a15 COLUMN failed_over FORMAT a11
SELECT instance_name , host_name , NULL AS failover_type , NULL AS failover_method , NULL AS failed_over FROM v$instance UNION SELECT NULL , NULL , failover_type , failover_method , failed_over FROM v$session;
Thanks! -- --Original Message-- -- From: oracle-l-bounce@(protected) [mailto:oracle-l-bounce@(protected)] On Behalf Of BERNHARD Nora Sent: Wednesday, July 13, 2005 11:37 PM To: oracle-l@(protected) Subject: TAF
Hello all,
I just set up a new database on a 2-node 10g RAC. There are already some other databases installed and running, with working failover. For the new db the TAF does not work. As far as I understand (newbie :( ), the following entries in tnsnames.ora should suffice:
My testing procedure (instructions from an Oracle support employee) is as follows: I open two separate sqlplus sessions to the same instance, in either of them I execute a long select, followed by a shutdown abort in the other session. The expected result is the select to continue running and finish gracefully on the second instance - but instead I receive "ORA-03113 (See ORA-03113.ora-code.com): end-of -file on communication channel".
