I have a question about write ahead protocol of Oracle.

If the *current* redo log is lost then
- Will Oracle halt? I think yes.

What happens if checkpoint crashes in the middle and the corresponding redo logs(those containing changes for checkpoint) are lost. As far as I understand the database could not be recovered without restore from backup.


Those statements are kind of supported by Oracle docs at http://download-east.oracle.com/docs/cd/B19306_01/backup.102/b14191/recoscen008.htm#sthref1872

Except that at some point in this doc Oracle states that:
...
If the group is Current - Attempt to clear the log; if impossible, then you must restore a backup and perform incomplete recovery up to the most recent available redo log.
...

But bellow Oracle states that:
... The current log is the one LGWR is currently writing to. If a LGWR I/O fails, then LGWR terminates and the instance crashes ...
(which is supported by my practical experience)

Isn?t this a contradiction?




Fyrirvari/Disclaimer
http://www.landsbanki.is/disclaimer
--
http://www.freelists.org/webpage/oracle-l