Not perfectly on topic, but if RAC is in scope, I always like to ask how crash recovery works in a cluster db. Particularly, when does it happen and what processes on which nodes are involved. This is also a good question to help determine who installed RAC once (and put it in their resume) versus those that have some understanding about how it works.
I don't think it is at all "nasty" to question applicants for their knowledge. They don't necessarily have to answer all correctly to pass the interview, though.
DanOn Feb 1, 2009, at 4:20 PM, Tony Adolph <tony.adolph.dba@xxxxxxxxx> wrote:
How about some example problems:How do you recover from a lost/corrupt SYSTEM / UNDO / other datafile. Or online redo log.What are your option if during a recovery you find that you have a missing / corrupt archived redo log.After all, its all well and good good knowing the theory, but its getting the DB backup with no / minimal data loss and asap (or within SLA) that's important.
-- //www.freelists.org/webpage/oracle-l