What is a checkpoint? List the operations to be performed by the system when a checkpoint is to be taken. What does the recovery system do if there is a crash?
There is large number of on-line database system exist so there could be hundreds of transactions handled per minute. The log for this kind of database consists of a very large volume of information. A scheme known as checkpoint is employed to limit the volume of log information that has to be handled and processed in the event of a system failure including the loss of volatile information. The checkpoint method is an extra component of the logging scheme.
A checkpoint operation, carried out periodically, copies log information onto stable storage and put the checkpoint marker at that time in the log. For every transaction active at checkpoint their identifiers and their database modification actions. Which at that time are reflected just only in the database buffers, will be propagated to the suitable storage.
Determine if there is a crash, it scans the log in a backward direction from the time of system crash. Basically, all the transaction is added to the undo list. Afterward, if it finds out that a transaction in the undo list has committed, that transaction is removed from the undo list and placed in the redo list. The redo list consists of all the transactions that have to be redone. All transactions which were committed before the checkpoint time require not be considered for the recovery operation. Hence, the overhead of the recovery from a system crash is reduced. Though, in a system that employs the transaction paradigm.
Checkpoint is a technique to minimize the search of the log and the amount of undo and redo needed recovering from a system failure along with loss of volatile storage.