Based on the manual ORACLE8i Backup and Recovery Guide Rel 8.15 Feb 1999 (pages1-4, 1-20, 1-23, 14-33), whenever you perform an incomplete recovery, you must reset the redo logs. My question is :

When I perform an tablespace point-in-time recover(TSPITR), do I have to reset the redo logs as TSPITR seems to be a incomplete recovery if it is not recovered to the current time
point ? If we do not need to do resetting the redo logs, how does Oracle maintain the redo logs? Does Oracle removes/clears all the change logs happened on that tablespace from the recovered time point to the current time point in the redo log files?

I would also like to know if I have to shutdown and reopen the database in order to reset the redo logs.

Another question is: Is the redo logs a database level or instance level? For example, for the parallel server, does each node/instance have its own redo logs, or the parallel server has only one redo logs that covers all nodes?

Thanks very much for your help.