Flashback Writer (RVWR) Behavior With I/O Errors

We know from very older verson of Oracle DB that Operation of LGWR
is so critical that if it unable to flush redo buffer then instance get
crashed this hapend even for Archiver process if DB is running in
archivelog mode.

now we have flashback feature from 10g onwards, so does this complications is with RVWR process also? Yes.

When
flashback is enabled or when there are guaranteed restore points, the
background process RVWR writes flashback data to flashback database
logs in the flash recovery area. If RVWR encounters an I/O error, the
following behavior is expected:

• If there are any guaranteed restore points defined, the instance will crash when RVWR encounters I/O errors.
• If no guaranteed restore points are defined, the instance will not crash when RVWR encounters I/O errors. There are 2 cases:
  • On
    a primary database, Oracle automatically disables flashback database
    while the database is open. All existing transactions and queries will
    proceed unaffected. This behavior is expected for both single instance
    and RAC.
  • On a physical or logical standby, RVWR will
    appear to be hung, retrying the I/O periodically. This may eventually
    hang the logical standby or the managed recovery of physical standby.
    (Oracle does not crash the standby instance because it does not want to
    crash the primary database in turn in max protection mode.) To resolve
    the hang, a DBA can either perform a SHUTDOWN ABORT or issue ALTER DATABASE FLASHBACK OFF.

Advertisements