Monday, April 16, 2007

Oracle9i DBA -- Database: Fundamentals I (7)

Maintaining Redo Log Files


  • The minimum size of a online redo log file is 50 KB. The maximum size is specific to the OS.
  • Redo log files are used only for recovery. Rollback segments provide the database with a read consistency method.
  • Archiving can be accomplished automatically by ARCn and manually through SQL statements. They can be multiplexed.
  • FAST_START_MTTR_TARGET parameter 700 : Recovery from an instance failure will not take more than 700 seconds. The number of buffers being written by DBWn is determined by the FAST_START_MTTR_TARGET parameter, if specified.
  • Archived Redo Log files can be accomplished automatically by ARCn and the maximum number of ARCn processes in an instance is 10
  • If the ARCn cannot write to the archive destination, oracle will write a message to the alert file and all database operations are stopped. Operations are resumed after a successful writing operation on the archived log file.
  • V$LOGFILE the value of the STATUS column could be one of the following:

    O INVALID indicates that the file is inaccessible.

    O STALE indicates that contents of the file are incomplete.

    O DELETED indicates that the file is no longer used.

    O Blank indicates that the file is in use.

  • Automatic archiving can be set by the parameter LOG_ARCHIVE_START = TRUE which is available in V$PARAMTER c is used to automate ARCn process only when the database is already in the ARCHIVELOG mode.
  • The ALTER DATABASE ARCHIVELOG is used when database is in mount stage.
  • %S specifies the log sequence number with leading zero filled. Eg: arch_%s -> arch_0000001
  • The CKPT writes checkpoint information in control files and data file headers.
  • The ARCHIVELOG mode does not start the ARChiver process, rather it states that the database may use archive logs. Hence it ensures that the complete recovery is possible even if the redo logs are overwritten through a log switch.
  • LOG_CHECKPOINT_TO_TRACE = true means that the information about the log switches will be stored in alert_sid.ora file.

  • When the oracle server starts, it assigns the same sequence number to a group, that may contain multiple members and all of them have the same sequence number.

Read more!

No comments: