Corruption in a log file

Question

Problem Description: alter database open
ERROR at line 1:
ORA-00368: checksum error in redo log block
ORA-00353: log corruption near block 75 change 898387731 time 11/07/2009 17:15:29
ORA-00312: online log 2 thread 1: '/opt/oracle/archivelogs/online_logs/redo02.log'
Elapsed: 00:00:01.71
SQL> select CONTROLFILE_CHANGE#, ARCHIVELOG_CHANGE# from v$database;
CONTROLFILE_CHANGE# ARCHIVELOG_CHANGE#
          898387746          898387674
Elapsed: 00:00:00.00
SQL> select distinct checkpoint_change# from v$datafile_header;
CHECKPOINT_CHANGE#
         898385334
Elapsed: 00:00:00.01I'm wondering if I should clear the logfile, or restore and recover control file to change# 898385334 or a higher number. Please advise. I filled out a severity 1 request, but I haven't heard back. It has been nearly 3 hours. I'm thinking I need to restore and recover the control file, but I would like some expert advice.
If it matters:
Oracle 11.1.0.6.0
the corruption was caused by a full disk (Oracle completely filled one archivelog destination while another had 600GB+ free)
the backups were done with RMAN
a full restore and recover will take considerable time (probably a week or more by the time I get the data off of tape)
I have all the archivelogs
I have a recent control file backup
and I'd be happy to post any other information.

Answer 1

Flashback is not enabled. I think it was enabled when we were running this on a win32 platform, but it's not anymore.
I thought the logfiles were duplexed when the database was set up, but Unfortunately, logfiles aren't duplexed. It also turns out that the datafiles are fuzzy. They need recovery from the corrupted online redo log, probably past the corruption point. Once the database is online again, I'll duplex or multiplex the logfiles before letting the customer use it again.
The database is about 1.3TB. A full backup normally takes two days to complete, plus it would a few days to free up the space on the backup server. However, I have some much faster disk space now.
Oracle support has gotten back to me, but their support system is suffering from partial outages. We tried applying all the logfiles, then an open resetlogs, but it didn't work, due to datafiles being fuzzy. His recommendation was to do a full restore and recover, but that will probably take more than a week. Most of the data has been archived to tape, and I've been generating 100's of GB (in their bzip2 compressed backup form) of archivelogs working on restoring the database from a previous crash.
He said we could also try to force the database open, do a full export, create the database again, then import the data, and said it had a pretty good chance of working. Unless I come across some other option, we'll try that first, since by morning, I'll have somewhere around 1.8TB of extra disk space to hold the data.