Physical standby - nologging operation
I've got a physical standby (18.104.22.168). I created a test table and a nologging index on it (the underlying tablespace is NOT in a force logging mode). Then pushed some data. Then did a "select name, unrecoverable_change# from v$datafile;" on both primary and standby. Surprisingly, I didn't see any difference in the change# and the standby was infact caughtup. I, then opened the standby in readonly mode. The index does seem to exist there. Am not sure if it is getting populated though. I tried forcing a hint to use the index etc. But I can't tell if it's actually using it, since I can't explain plans on a read-only. I suppose I can run traces etc..but before I go down the path, can someone explain this phenomenon? What might be going on?
Re: Physical standby - nologging operation
The index is created and logged but the population of the index is not done and after recovery the blocks occupied by the index are marked as corrupted.
If you try to access the index, you will get an error.