Project Title: Implementation and Evaluation of Database Disaster Management System (Database shadowing)
Aim: The project aims to implement database recovery system in Oracle with Object Relational Database Management System. This will implement the shadow database system which is exactly having the live copies of the primary database
Deliverable: Implementing a mirror database using the database storage and recovery techniques. This application implements the shadow database in such a way that both primary database and the shadow database are in synchronisation.
- How effectively database mirroring handle the complex data with the help of object relational database management system.
Relational Database Management System (RDBMS) is the database that helps to organizing, storing and retrieving data, and also it eliminates data redundancy (Kanjilal, 2007). Object-oriented database system (OODBMS) is able to store complex data such as Objects, but it is slow to add SQL support to the system (Wong and Fong, 2000). There is a big gap between the Relational Database and Object –Oriented Database. That led to evolve Object Relational Database (ORDBMS) to fulfil the gap between Relational Database and Object Oriented database by implementing Object Oriented features in Relational Database (Wong and Fong, 2000). The database mirroring is an instance that wills continuously sends database’s transaction log records to the shadow database that is exist on the other machine (Talmage, 2005). The Database shadowing technique will prevent the data form the several threat categories.
- Data loss
- Data corruption
- Data inaccessibility
Data corruption and Data inaccessibility will cause the primary database to go down as soon as the primary database is stopped working the shadow database will starts it functioning. In this we should give similar name for both Primary database and the mirror database. In The following scenario we need database mirroring.
In an organization one or more servers or databases go down it leads to the failure transaction to eliminate this situation we need to develop a mirror database to avoid the data loss.
The project deals with the evaluation and implementation of Object Relational DBMS that involves the installation of required software Oracle database and Oracle JDeveloper.
Evaluation: This project will involve in implementation of database recovery techniques. I must start my project by finding right material. Database mirroring is a quite new era in database management system. To cope up with this project I must need to learn database storage and recovery techniques very well.
1: the shadow database provides the valuable protection of the data loss in addition to the regular backups.
2: If the Primary database is damaged or go down the shadow database is starts functioning in a little time.
3: The maintenance of the shadow database is very little and it uses the same process as the primary database.
Hardware: Computer system.
Software: Oracle database.
Data base can work with two or many more copies of database information simultaneously and this process is known as Database shadowing. The base copy is the master file, which is the source of data for all other duplicate files. The other copies can be other wise called as shadows. The shadows are updated whenever the master copy is updated; with a view to keep all the shadow copies up-to-date of the master. However, the shadows have no impact on response times under normal working conditions. Changes are executed on the master data and its shadows by using background threads, such that it doesn’t interrupt user functions.  This database shadowing plays a major role when the master data bank is lost. That is, when the master data bank is lost, the shadow can speedily take over control of the master and continue the operations immediately (assuming that the shadow is safe from disk collapse). This is because, renovating a shadow to a master is far more quickly done than reproducing from a back up print. This application is entirely hidden and is not visible; shadows can be altered to already present application also. 
Database shadowing in Oracle:
As we already know, data shadowing is the process of duplication of database information such that it might be useful at the time of failures in the system for recovery purposes. Disk shadowing is one of the processes through which data is updated simultaneously to all the shadows. This type of storage permits to load the data in shadow disks to primary disks or any other disk in case of system failure. Resilvering is a development in which the shadowing process is brought back to synchronizing state, once the disk is kept online again. This process facilitates a back up source in case if shadow or mirror set gets wrecked up. Thus, by proper use of resilvering procedure the set can be resynchronized. Mirror sets can be in any number.
The major advantages of disk shadowing in Oracle are:
- Facilitates fast and secure recovery of data in case of Disk failure or disk crash.
- Exhibits better performance as most operating systems supports volume shadowing which that can exhibit file I/O request to use the shadow set of files instead of the main set of files.
As a coin has both sides, there will be disadvantages also if there are advantages, the major disadvantage is, and this requires 400 or more giga bytes of storage for a 200GB disk farm.
Database Disaster Recovery management:
Recovery from crash is a critical factor in databases. When a system undergoes crash (disaster), care should be taken that it should be recovered completely (if not as much as possible)
Database: Database is a group of related storage stuff gathered together with redundancy which can serve many applications. Data is stored in a database and the single approach programs like adding, modifying, updating are applied on it.
Disaster: Disaster or failure is the situation where in, system doesn’t perform according to the requirements specified. This disaster can be of many kinds, of which three occur majorly. They are:
- Transactions or programs may get collapsed.
- Total system break down
- Mal function of hardware
Recovery: In highly reliable systems, data may also be required to cover the recovery mechanism. Recovery of data can be sorted into two ways:
- Data necessary to keep the current values
- To make renovation of possible previous values
The degree of recovery management varies from system to system. Some of the probable levels of recovery are:
- Correct position recovery
- Restoration to the past check point or the specified check point
- Specified previous state should be restored
- Valid state should be recuperated
- The recovery technique depends on the extent of damage for the system.
Some of the recovery techniques that can be implemented for disaster management are listed below:
- Salvation program: Affording to re-establish the system back to the valid position crash should be programmed after the crash. Salvation program is mainly used when all other programs fail to function. This program works the best in situations where buffers were lost during crash and striving to rebuild the lost data.
- Incremental dumping: This technique is one of the formal ways of restoration. This involves transferring of copied files to archive folder, once the files are modified or copied at regular intervals of time.
- Audit trail: The continual chain of files should be verified by recording them. This is one of the most favorable things for backing out of transactions.
- Differential files: To keep track of the changes made for the files, a separate document is maintained and recorded accordingly. These recordings are then merged with the main file.
- Backup/current version: Present files form the current version of the database. Files containing previous values form a consistent backup version.
- Multiple copies: During normal operation, keeping in view that any disaster may occur, multiple copies of each file are preserved. Thus, when a failure occurs, the consistent version can be obtained by comparing the versions.
- Careful replacement: careful replacement should be done while updating, with the original only being deleted after operation is complete.
Recovery techniques can be provided in a better way by combining the above stated techniques such that it would be helpful for different kinds of failures, some of them are like:
- How the data can be controlled and changes can be done
- The way how data can be structured
Oracle Disaster Recovery:
For any administrator, the very powerful tool that is mostly used is Oracle database. Irrespective of the organisational setup, either it may be Educational institution, business organisation or any other business organisation; database system has become a critical advantage and also a part and parcel of their business processing. Thus, this leads to great demand for Oracle Disaster Recovery management which is an important object for an administrator recognize.
As per the views of many Oracle administrators, it was known that Oracle Disaster Recovery occurs well before a disaster ever occurs. Recovery becomes very critical when physical hard drive failure occurs.
Having a healthy backup system, is one of the best ways to tackle Oracle Disaster Recovery plan. This is regarded to be the highest priority when listing in descending order. This implies that remote back up system is a must kind of thing for any system. If not, try to have a secondary back up using tapes. If u have a secondary back up then try to ensure that swapping them at least once for a week.
Need for professional help:
If is obvious that, every administrator cannot get 100% efficient back up system and if any hard ware problems occur, it needs the help of the Oracle data recovery expert who can rectify it.
In cases like this, you will definitely require a professional data recovery service such as the Hard Drive Recovery Group. This fact is, even the most savvy IT professional likely does not have access to the tools, or the Clean Room facilities necessary for professional data recovery, much less the mission critical nature of Oracle Disaster Recovery.