Log based recovery, Database Management System

Log based recovery:Let us first describe the term transaction log in the context of DBMS. A transaction log is a record in DBMS that remains track of all the transactions of a database system that update any data values in the database. A log haves the following information about a transaction:

  • A transaction start marker
  • The transaction identification: The transaction id, user id or terminal id etc.
  • The operations being performed by the transaction such as delete, update, insert.
  • The objects or data items that are affected by the transaction as well as name of the table, row number and column number.
  • The previous or before values (also known as UNDO values) and changed or after values (also called REDO values) of the data items that have been updated.
  • A pointer to the next transaction log record, if required.
  • The COMMIT marker of the transaction.

In a database system various transactions run simultaneously. When a transaction commits, the data buffers used by it require not be written back to the physical database stored on the secondary storage as these buffers may be used by various other transactions that have not yet committed.  On the other hand, some of the data buffers that may have updates by various uncommitted transactions may be forced back to the physical database, as they is no longer being used by the database.  So the transaction log helps in remembering which transaction did which alters.  Therefore the system knows exactly how to divide the changes made by transactions that have already committed from those alters that is made by the transactions that did not yet commit.  Any operation such as start transaction, insert /delete/update and end transaction (commit), adds information to the log having the transaction identifier and enough information to redo or undo the changes.

But how do we recover using log? Let us show this with the help of an example having three concurrent transactions that are active on ACCOUNTS table as:

 

Transaction T1

Transaction T2

Transaction T3

Read X

Read A

Read Z

Subtract 100

Add 200

Subtract 500

Write X

Write A

Write Z

Read Y

 

Add 100

Write Y

                                                Figure: The sample transactions

Suppose that these transactions have the following log file (hypothetical) at a point:

Transaction Begin Marker

Transaction

Id

Operation on ACCOUNTS table

UNDO values (assumed)

REDO

values

Transaction Commit Marker

Y

T1

Sub on X Add on Y

500

800

400

Not done yet

N

Y

T2

Add on A

1000

1200

N

Y

T3

Sub on Z

900

400

Y

                                 Figure: A sample (hypothetical) Transaction log

Now suppose at this point of time a failure occurs, then how the recovery of the database will be completed on restart.

Values

Initial

Just before the failure

Operation

Required for recovery

Recovered Database Values

X

500

400 (assuming update has been done in physical database also)

UNDO

500

Y

800

800

UNDO

800

A

1000

1000 (assuming update has not been done in physical database)

UNDO

1000

Z

900

900 (assuming update has not been done in physical database)

REDO

400

 

Posted Date: 3/12/2013 5:18:06 AM | Location : United States







Related Discussions:- Log based recovery, Assignment Help, Ask Question on Log based recovery, Get Answer, Expert's Help, Log based recovery Discussions

Write discussion on Log based recovery
Your posts are moderated
Related Questions
Object-based Models: Use objects as key data representation components . Examples:-   Entity-Relationship Model: It is a collection of real world objects known as entitie

Modify a sequence: ALTER SEQUENCE SEQSS INCREMENT 2 MAXVALUE 3000;   Removing a sequence: DROP SEQUENCE SEQSS;

Equi Join: A join in which the joining condition is based on equality among values in the common columns. Common columns show (redundantly) in the result table. Let us Consider th

Sub query in the UPDATE command : Example :  //Double  the  commission  for  employees,  who  have  got  at  least  2 increments. UPDATE EMP SET COMM = COMM * 2 WHERE

Create 1 experiment to demonstrate an aspect of query optimisation by using access-path structures (indexes), decomposition, or denormalisation. Once satisfied with your experiment

Question: Consider the following database relations for a textbook ordering system used by a college bookshop: Book (b-copy#, bname, ISBN#, author, price, subject) Order (s#

Consider the relations EMPLOYEE(emp#, name) ASSIGNED_TO(project#, emp#) PROJECT(project#, project_name, chief) Express the subsequent queries in Relational Algebra (i) Ge

Define Radix conversion method  One clever way to transform binary numbers to BCD notation (binary-coded decimal) is the "double dabble algorithm". It can be adapted to transfo

Explain the benefit of OOM. Main benefits of object oriented modeling are development of system become fast, quality of the system get improve. It gives freedom of use of alrea

What do you mean by persistence? How you will make your data persistent? Persistent data is the data, which has a longer lifetime than program which has created it. Enabling th