Oracle Data Guard Architecture

Data is one, if not the main value of the business, data security is important for the database administrator, this includes timely backup and recovery. I'm not going to lie, but cost plays a big part in the decision to implement a particular solution. Oracle Data Guard can be overwhelming for many businesses. Having a solution that takes minutes will not be cheap, and many small companies do. that I work for are very happy to be able to copy the redo logs to a static machine and create a script to add those logs, this may not be the best solution but it works. This has disadvantages because it is a scalable solution that is difficult to manage or transfer to other database administrators. Oracle's Data Guard is well-tested and supported, but it is not cheap it is a solution that is almost universal in all companies and that all data protection administrators should be able to understand quickly when they start. new business.

In this series, I cover the architecture, installation, and configuration of Oracle Data Guard 11g, and I will cover operational planning and how to implement solutions to ensure data integrity. I also explain how to use unused storage databases, thus getting the most out of your system. I will also cover the new Enterprise Manager and how it fits into the database, most management tools use GUI interfaces, so there is no escaping it, the way I will learn both because the GUI will run the command line and the experience. and a good understanding of the law will help you find and solve difficult issues.

My advice is the more you practice, the more you do it the better you will do it and it will become second nature, in today's world, there are many good solutions (VMware, HyperV, etc.) that you can. Configure and do whatever you want, even some of these virtual solutions are free (VMware ESXi), all you need is a spare server or PC (see my virtual solution), extra RAM, and disk space. I can set up a solution similar to what I have on the company site, test and learn how the solution works, and get an idea of ​​what I need to do if I need to change, I can make and record the products before time. this happens, in some cases, I have worked in a company and I will never change, but I am always happy that I am doing something and have the book in hand. Okay, I'm done rambling, let's start by learning Oracle Data Guard. 

Data Guard Overview

Data security is basically a recovery and put a job because as you know, recovery is the information needed to recover the data security business. A production database called the primary database forwards recovery to one or more independent replicas called standby databases. The database is in continuous recovery mode, supporting and applying recovery to maintain relationships with the primary database. The backup database will automatically recover if it is temporarily disconnected from the primary database due to power outages, network problems, etc. 

     A Dataguard configuration consists of one primary database and one or more standby databases. The standby databases will be always in sync with the primary database. If the primary production database is unavailable due to a planned or unplanned outage, then we can switch the production to the standby database, making the application less affected by the outage.

The main features of Oracle Data Guard are

1. High Availability

2. Data protection

3. Disaster Recovery

    

                                                FIGURE 1-01. Typical Data Guard Configuration

The diagram above shows the overview of Data Guard First, the recovery service transfers the recovery data from the primary to the standby server after it is created. Second, users enter recovery data and update the standby database file. Third, independent of the database, the author of the database. This process updates the primary database files and eventually, Data Guard will automatically restore the standby database following a power or network failure using the recovery data stored in the primary database.

.If you haven't read my Oracle process on rollback, this is a rollback, Roll Index (also called rollback) is a group of changes, each of which is a description of changes made in one block. . and data protection. Restore records contain all the necessary information to restore the changes made to the database. During recovery, the database will read the change vectors in the redo record and apply changes to the affected blocks.

Redo logs are stored cyclically in the SGA's redo buffer, and the log write process (LGWR) is a backup process that manages redo buffer management. LGWR writes log entries to batch files (internet log files) at specific intervals to provide access and security, LGWR writes.

• A commit record - Each time a transaction is made, LGWR writes a log of the transaction from storage to a log file and submits a System Change Number (SCN). Only after this process is completed is the transaction confirmed.
• Redo log buffers - If the redo log is 1/3 full or if 3 seconds have passed since LGWR last wrote to the log file, all entries in the backup folder will be written to the log file. This means that a redo record can be written to the log file before the transaction is made and, if necessary, the media recovery will repeat the changes using the correction which is also part of the input recovery.

FIGURE 1-02. Data Guard Redo Transport

Note that LGWR can use "group" in log files. Essentially, all pending (uncommitted) transaction redo entries can be written to disk in a single operation, thereby reducing I/O. Even if the database cache is not written to disk, Oracle guarantees that no transaction will be lost due to the successful change log and redo log.

The Redo Data Transport Backup service coordinates the transfer of recovery from the primary database to the standby database. At the same time as the LGWR performs the recovery operation, a separate database system called the Log Network Server (LNS) reads from the recovery storage in the SGA and forwards the recovery to Oracle Net Services by forwarding to database alert, it is possible. restore data directly to nine storage databases, you can also use Oracle RAC but not all of them need to be RAC configurations. The Remote File Server (RFS) system receives the redo from the LNS and writes it to a system file called a fallback redo log (SRL) file. The LNS system supports two modes: synchronous and asynchronous.

A Dataguard configuration consists of one primary database and one or more standby databases. The standby databases will be always in sync with the primary database. If the primary production database is unavailable due to a planned or unplanned outage, then we can switch the production to a standby database, making the application less affected by the outage.

Data Guard Management

PROTECTION MODES:

It can have 3 protection modes.

1. MAXIMUM PROTECTION : 

This type provides maximum protection. It guarantees zero data loss. In this mode, the redo/transaction data must be written to both the primary redo log and the backup redo log. For some reason (mostly N/W editions), if it doesn't write while sleeping, Primary will stop.

2. MAXIMUM AVAILABILITY : 

It provides the highest possible data protection without affecting the availability of the main data storage. Transactions are not allowed until all redo data is written into the online redo log and propagated to at least one synchronized secondary database. If for some reason the primary database cannot propagate its recovery flow to a secondary database, the primary database will NOT shut down and will run in near maximum performance mode until until the problems are resolved.

3. MAXIMUM PERFORMANCE : 

This is the default mode. From this safety, a business is started once data recovery data to restore the transaction is written to the local (online) redo log.