You encounter the following errors being signaled in the alert log periodically:
ORA-01242: data file suffered media failure: database in NOARCHIVELOG mode
ORA-01122: database file %d failed verification check
ORA-01110: data file %d: '%s'
ORA-01208: data file is an old version - not accessing current version
The data file specified however stays on-line and available so no apparent corruption exists, and no I/O errors are encountered on the underlying disk.
Cause
This problem can occur when running a database in NOARCHIVELOG mode due to heavy disk I/O usage such that I/O performance is impacted, for example when redo log switches are occurring too frequently, such that the control file gets updated with the latest checkpoint scn, but the data file is then accessed prior to its header being updated with the same checkpoint scn information, and so Oracle reports the ORA-1208 error as the data file is older than listed in the control file. Subsequently the data file header is updated and so can be accessed normally with no corruption having occurred.
Solution
The solution here is to resolve the underlying disk I/O performance issue. If this is occurring due to redo log switches occurring too quickly, i.e. a few seconds apart, then refer to Note:147468.1 on how to tune checkpoint operations and in particular section 4 relating to correct redo log file sizing (see Note:1038851.6) to resolve the problem.
ORA-01242: data file suffered media failure: database in NOARCHIVELOG mode
ORA-01122: database file %d failed verification check
ORA-01110: data file %d: '%s'
ORA-01208: data file is an old version - not accessing current version
The data file specified however stays on-line and available so no apparent corruption exists, and no I/O errors are encountered on the underlying disk.
Cause
This problem can occur when running a database in NOARCHIVELOG mode due to heavy disk I/O usage such that I/O performance is impacted, for example when redo log switches are occurring too frequently, such that the control file gets updated with the latest checkpoint scn, but the data file is then accessed prior to its header being updated with the same checkpoint scn information, and so Oracle reports the ORA-1208 error as the data file is older than listed in the control file. Subsequently the data file header is updated and so can be accessed normally with no corruption having occurred.
Solution
The solution here is to resolve the underlying disk I/O performance issue. If this is occurring due to redo log switches occurring too quickly, i.e. a few seconds apart, then refer to Note:147468.1 on how to tune checkpoint operations and in particular section 4 relating to correct redo log file sizing (see Note:1038851.6) to resolve the problem.
