Today, I lost all my redologs and obviously,I made an incomplete recovery this way:
As I didn't had the redo log files(including the current log file), I faked up the redo logfiles using 'mkfile'command:
restored the backup datafiles and control files
mkfile 10485760 redo01.log
mkfile 10485760 redo02.log
mkfile 10485760 redo03.log
recover database until cancel using backup controlfile
alter database open resetlogs;
I was under the impression that I might need to use _allow_redolog_corruption = TRUE to recover from this type of disaster but just faking the redo logs helped.
I am just wondering why oracle did not complain about the missing current redolog and about the 'NEW' redolog files which are not in the 'ORACLE' format.
"Congrats"... You will pick the lottery numbers...
Your redo logs did not have any information in them that was necessary for recovery (i.e. The last checkpoint # was contained in an archive redo log). Oracle just needs to start overwriting the redo logs when the database is opened, it does not verify their structure.