Online Distributed Database Transaction And Recovery Techniques
When it comes to online, database failure can be disturbing. Losing valuable data can be depressing because sometimes data is lost accidentally when we least expect this to happen. But data loss prevention is one of the most vital issues involved in managing database systems. Data can be lost due to viruses, hardware failures, incorrect use of delete and update statements, and disasters, such as fire or flood, software bugs.
Implementing a recovery strategy for your databases can help avoid data loss. Here are some of the restoration techniques
Transaction undo or rollback
This method can be applied to do away with the undesirable effects of faulty transactions. Erroneous transactions include all operations that have altered the database into unwanted state and the operations that have used values written by the defective transactions. Therefore, transaction undo applies in situations where soft failures can result in the discrepancy of the database. Transaction redo is implemented to recuperate to a steady transaction state.
This is when recording is done on the folder from the hardware shared data area. Checkpointing eliminates the redoing of transactions performed earlier. Periodical checkpointing reduces the recuperation procedure. There are two kinds of checkpointing:
Fuzzy: This is where all active transactions are recorded in the log at the time of checkpoint. Therefore, in the case of a power blackout or power failure, only those transactions that were active during checkpoint and later are processed. All other transactions, done before the checkpoint, are written to the disk, and there is no need for redoing.
Consistent: This checkpointing method generates a dependable representation of the database at the checkpoint. In the event of recovery, right side operations are the only ones undone and redone.
Database backup and Transaction log
This technique is useful in case of a hard crash or disk malfunction that may lead to overall loss of the database. A new drive is set so as to help recuperate from this disk fault and reinstate the operating system. And this is where the transaction log and database backing come in to complete the database recovery process.
The method includes:
l Redoing and writing of activity in the before-commit list and the commit list into the commit list in the operation log.
l Undoing and writing the transactions on the failed list and busy list into the terminate list in the operation record.
If you are concerned about your database and the transactions, backups can be vital in the event of any data loss.