9 Configuring Oracle GoldenGate for Active-Active High Availability
Topics:
- Overview of an Active-Active Configuration
- Considerations for an Active-Active Configuration
- Preventing Data Looping
- Managing Conflicts
- Additional Information
- Creating an Active-Active Configuration
Parent topic: Administering Oracle GoldenGate Classic Architecture
9.1 Overview of an Active-Active Configuration
Oracle GoldenGate supports an active-active bi-directional configuration, where there are two systems with identical sets of data that can be changed by application users on either system. Oracle GoldenGate replicates transactional data changes from each database to the other to keep both sets of data current.
In a bi-directional configuration, there is a complete set of active Oracle GoldenGate processes on each system. Data captured by an Extract process on one system is propagated to the other system, where it is applied by a local Replicat process.
This configuration supports load sharing. It can be used for disaster tolerance if the business applications are identical on any two peers. Bi-directional synchronization is supported for all database types that are supported by Oracle GoldenGate.
Oracle GoldenGate supports active-active configurations for:
-
DB2 on z/OS, LUW, and IBM i
-
MySQL
-
Oracle
-
SQL Server
Oracle GoldenGate supports DDL replication in an Oracle active-active configuration. DDL support is available for Oracle Database and MySQL (like-to-like configuration) databases.
9.2 Considerations for an Active-Active Configuration
The following considerations apply in an active-active configuration. In addition, review the Oracle GoldenGate installation and configuration document for your type of database to see if there are any other limitations or requirements to support a bi-directional configuration.
9.2.1 TRUNCATES
Bi-directional replication of TRUNCATES
is not supported, but you can configure these operations to be replicated in one direction, while data is replicated in both directions. To replicate TRUNCATES
(if supported by Oracle GoldenGate for the database) in an active-active configuration, the TRUNCATES
must originate only from one database, and only from the same database each time.
Configure the environment as follows:
-
Configure all database roles so that they cannot execute
TRUNCATE
from any database other than the one that is designated for this purpose. -
On the system where
TRUNCATE
will be permitted, configure the Extract and Replicat parameter files to contain theGETTRUNCATES
parameter. -
On the other system, configure the Extract and Replicat parameter files to contain the
IGNORETRUNCATES
parameter. NoTRUNCATES
should be performed on this system by applications that are part of the Oracle GoldenGate configuration.
Parent topic: Considerations for an Active-Active Configuration
9.2.2 Application Design
When using Active-Active replication, the time zones must be the same on both systems so that timestamp-based conflict resolution and detection can operate.
Active-active replication is not recommended for use with commercially available packaged business applications, unless the application is designed to support it. Among the obstacles that these applications present are:
-
Packaged applications might contain objects and data types that are not supported by Oracle GoldenGate.
-
They might perform automatic DML operations that you cannot control, but which will be replicated by Oracle GoldenGate and cause conflicts when applied by Replicat.
-
You probably cannot control the data structures to make modifications that are required for active-active replication.
Parent topic: Considerations for an Active-Active Configuration
9.2.3 Keys
For accurate detection of conflicts, all records must have a unique, not-null identifier. If possible, create a primary key. If that is not possible, use a unique key or create a substitute key with a KEYCOLS
option of the MAP
and TABLE
parameters. In the absence of a unique identifier, Oracle GoldenGate uses all of the columns that are valid in a WHERE
clause, but this will degrade performance if the table contains numerous columns.
To maintain data integrity and prevent errors, the following must be true of the key that you use for any given table:
-
contain the same columns in all of the databases where that table resides.
-
contain the same values in each set of corresponding rows across the databases.
Parent topic: Considerations for an Active-Active Configuration
9.2.4 Triggers and Cascaded Deletes
Triggers and ON DELETE CASCADE
constraints generate DML operations that can be replicated by Oracle GoldenGate. To prevent the local DML from conflicting with the replicated DML from these operations, do the following:
-
Modify triggers to ignore DML operations that are applied by Replicat. If the target is an Oracle database, Replicat handles triggers without any additional configuration when in integrated mode. Parameter options are available for a nonintegrated Replicat for Oracle. See Diabling Triggers and Referential Cascade Constraints on Target Tables in Using Oracle GoldenGate for Oracle Database.
-
Disable
ON DELETE CASCADE
constraints and use a trigger on the parent table to perform the required delete(s) to the child tables. Create it as aBEFORE
trigger so that the child tables are deleted before the delete operation is performed on the parent table. This reverses the logical order of a cascaded delete but is necessary so that the operations are replicated in the correct order to prevent "table not found" errors on the target.Note:
For MySQL targets, cascade delete queries result in the deletion of the child of the parent operation.
Note:
For Oracle Database targets, if Replicat is in integrated mode, constraints are handled automatically without special configuration.
Parent topic: Considerations for an Active-Active Configuration
9.2.5 Database-Generated Values
Do not replicate database-generated sequential values, such as Oracle sequences, in a bi-directional configuration. The range of values must be different on each system, with no chance of overlap. For example, in a two-database environment, you can have one server generate even values, and the other odd. For an n-server environment, start each key at a different value and increment the values by the number of servers in the environment. This method may not be available to all types of applications or databases. If the application permits, you can add a location identifier to the value to enforce uniqueness.
Parent topic: Considerations for an Active-Active Configuration
9.2.6 Database Configuration
One of the databases must be designated as the trusted source. This is the primary database and its host system from which the other database is derived in the initial synchronization phase and in any subsequent resynchronizations that become necessary. Maintain frequent backups of the trusted source data.
Parent topic: Considerations for an Active-Active Configuration
9.3 Preventing Data Looping
In a bidirectional configuration, SQL changes that are replicated from one system to another must be prevented from being replicated back to the first system. Otherwise, it moves back and forth in an endless loop, as in this example:
-
A user application updates a row on system A.
-
Extract extracts the row on system A and sends it to system B.
-
Replicat updates the row on system B.
-
Extract extracts the row on system B and sends it back to system A.
-
The row is applied on system A (for the second time).
-
This loop continues endlessly.
To prevent data loopback, you may need to provide instructions that:
-
prevent the capture of SQL operations that are generated by Replicat, but enable the capture of SQL operations that are generated by business applications if they contain objects that are specified in the Extract parameter file.
-
identify local Replicat transactions, in order for the Extract process to ignore them.
9.3.1 Preventing the Capture of Replicat Operations
Depending on which database you are using, you may or may not need to provide explicit instructions to prevent the capture of Replicat operations.
- Preventing the Capture of Replicat Transactions (Oracle)
- Preventing Capture of Replicat Transactions (Other Databases)
Parent topic: Preventing Data Looping
9.3.1.1 Preventing the Capture of Replicat Transactions (Oracle)
To prevent the capture of SQL that is applied by Replicat to an Oracle database, there are different options depending on the Extract capture mode:
-
When Extract is in classic or integrated capture mode, use the
TRANLOGOPTIONS
parameter with theEXCLUDETAG
tag
option. This parameter directs the Extract process to ignore transactions that are tagged with the specified redo tag. See Identifying Replicat Transactions to set the tag value. -
When Extract is in classic capture mode, use the Extract
TRANLOGOPTIONS
parameter with theEXCLUDEUSER
orEXCLUDEUSERID
option to exclude the user name or ID that is used by Replicat to apply the DDL and DML transactions. MultipleEXCLUDEUSER
statements can be used. The specified user is subject to the rules of theGETREPLICATES
orIGNOREREPLICATES
parameter. See Preventing Capture of Replicat Transactions (Other Databases) for more information.
Parent topic: Preventing the Capture of Replicat Operations
9.3.1.2 Preventing Capture of Replicat Transactions (Other Databases)
To prevent the capture of SQL that is applied by Replicat to other database types (including Oracle, if Extract operates in classic capture mode), use the following parameters:
-
GETAPPLOPS | IGNOREAPPLOPS
: Controls whether or not data operations (DML) produced by business applications except Replicat are included in the content that Extract writes to a specific trail or file. -
GETREPLICATES | IGNOREREPLICATES
: Controls whether or not DML operations produced by Replicat are included in the content that Extract writes to a specific trail or file.
Parent topic: Preventing the Capture of Replicat Operations
9.3.2 Identifying Replicat Transactions
To configure Extract to identify Replicat transactions, follow the instructions for the database from which Extract will capture data.
9.3.2.1 DB2 z/OS, DB2 LUW, and DB2 for i
Identify the Replicat user name by using the following parameter statement in the Extract parameter file.
TRANLOGOPTIONS EXCLUDEUSER user
This parameter statement marks all DDL and DML transactions that are generated by this user as Replicat transactions. The user name is included in the transaction record that is read by Extract.
Parent topic: Identifying Replicat Transactions
9.3.2.2 MySQL
Identify the name of the Replicat checkpoint table by using the following parameter statement in the Extract parameter file.
TRANLOGOPTIONS FILTERTABLE table_name
Replicat writes a checkpoint to the checkpoint table at the end of each of its transactions as part of its checkpoint procedure. (This is the table that is created with the ADD CHECKPOINTTABLE
command.) Because every Replicat transaction includes a write to this table, it can be used to identify Replicat transactions in a bidirectional configuration. FILTERTABLE
identifies the name of the checkpoint table, so that Extract ignores transactions that contain any operations on it.
Parent topic: Identifying Replicat Transactions
9.3.2.3 Oracle
There are multiple ways to identify Replicat transaction in an Oracle environment. When Replicat is in classic or integrated mode, you use the following parameters:
-
Use
DBOPTIONS
with theSETTAG
option in the Replicat parameter file. Replicat tags the transactions being applied with the specified value, which identifies those transactions in the redo stream. The defaultSETTAG
value is00
. Valid values are a single TAG value consisting of hexadecimal digits.
-
Use the
TRANLOGOPTIONS
parameter with theEXCLUDETAG
option in the Extract parameter file. The logmining server associated with that Extract excludes redo that is tagged with theSETTAG
value.The following shows how
SETTAG
can be set in the Replicat parameter file:DBOPTIONS SETTAG 0935
The following shows how
EXCLUDETAG
can be set in the Extract parameter file:TRANLOGOPTIONS EXCLUDETAG 0935
If you are excluding multiple tags, each must have a separate
TRANLOGOPTIONS EXCLUDETAG
statement specified.
You can also use the transaction name or userid of the Replicat user to identify Replicat transactions. You can choose which of these to ignore when you configure Extract. See Preventing the Capture of Replicat Transactions (Oracle).
For more information, see Reference for Oracle GoldenGate.
Parent topic: Identifying Replicat Transactions
9.3.2.4 SQL Server
(CDC Extract) Identify the name of the Replicat checkpoint table by using the following parameter statement in the Extract parameter file and ensure that the Replicat checkpoint table has been enabled for supplemental logging with the ADD TRANDATA
command.
TRANLOGOPTIONS FILTERTABLE table_name
Replicat writes a checkpoint to the checkpoint table at the end of each of its transactions as part of its checkpoint procedure. (This is the table that is created with the ADD CHECKPOINTTABLE
command). Because every Replicat transaction includes a write to this table, it can be used to identify Replicat transactions in a bi-directional configuration. FILTERTABLE
identifies the name of the checkpoint table, so that Extract ignores transactions that contain any operations on it.
(Classic Extract) By default, Extract ignores the Replicat's transactions, however, if you modify the Replicat's transaction name with the DBOPTIONS TRANSNAME
parameter, then you must exclude those transactions by using the following parameter statement in the Extract parameter file.
TRANLOGOPTIONS EXCLUDETRANS transaction_name
This parameter statement is only required if the Replicat transaction name is set to something other than the default of ggs_repl
.
Parent topic: Identifying Replicat Transactions
9.3.3 Replicating DDL in a Bi-directional Configuration
Additional consideration must be taken when replicating DDL bi-directionally, currently only supported for Oracle database. For more information, see Managing the DDL Replication Environment Using Oracle GoldenGate for Oracle Database.
Parent topic: Preventing Data Looping
9.4 Managing Conflicts
Uniform conflict-resolution procedures must be in place on all systems in an active-active configuration. Conflicts should be identified immediately and handled with as much automation as possible; however, different business applications will present their own unique set of requirements in this area.
Because Oracle GoldenGate is an asynchronous solution, conflicts can occur when modifications are made to identical sets of data on separate systems at (or almost at) the same time. Conflicts occur when the timing of simultaneous changes results in one of these out-of-sync conditions:
-
A uniqueness conflict occurs when Replicat applies an insert or update operation that violates a uniqueness integrity constraint, such as a
PRIMARY KEY
orUNIQUE
constraint. An example of this conflict type is when two transactions originate from two different databases, and each one inserts a row into a table with the same primary key value. -
An update conflict occurs when Replicat applies an update that conflicts with another update to the same row. Update conflicts happen when two transactions that originate from different databases update the same row at nearly the same time. Replicat detects an update conflict when there is a difference between the old values (the before values) that are stored in the trail record and the current values of the same row in the target database.
-
A delete conflict occurs when two transactions originate at different databases, and one deletes a row while the other updates or deletes the same row. In this case, the row does not exist to be either updated or deleted. Replicat cannot find the row because the primary key does not exist.
For example, UserA on DatabaseA updates a row, and UserB on DatabaseB updates the same row. If UserB's transaction occurs before UserA's transaction is synchronized to DatabaseB, there will be a conflict on the replicated transaction.
A more complicated example involves three databases and illustrates a more complex ordering conflict. Assume three databases A, B, and C. Suppose a user inserts a row at database A, which is then replicated to database B. Another user then modifies the row at database B, and the row modification is replicated to database C. If the row modification from B arrives at database C before the row insert from database A, C will detect a conflict.
Where possible, try to minimize or eliminate any chance of conflict. Some ways to do so are:
-
Configure the applications to restrict which columns can be modified in each database. For example, you could limit access based on geographical area, such as by allowing different sales regions to modify only the records of their own customers. As another example, you could allow a customer service application on one database to modify only the
NAME
andADDRESS
columns of a customer table, while allowing a financial application on another database to modify only theBALANCE
column. In each of those cases, there cannot be a conflict caused by concurrent updates to the same record. -
Keep synchronization latency low. If UserA on DatabaseA and UserB on DatabaseB both update the same rows at about the same time, and UserA's transaction gets replicated to the target row before UserB's transaction is completed, conflict is avoided. See Tuning the Performance of Oracle GoldenGate for suggestions on improving the performance of the Oracle GoldenGate processes.
To avoid conflicts, replication latency must be kept as low as possible. When conflicts are unavoidable, they must be identified immediately and resolved with as much automation as possible, either through the Oracle GoldenGate Conflict Detection and Resolution (CDR) feature, or through methods developed on your own. Custom methods can be integrated into Oracle GoldenGate processing through the SQLEXEC
and user exit functionality. See Configuring Conflict Detection and Resolution for more information about using Oracle GoldenGate to handle conflicts.
For Oracle database, the automatic Conflict Detection Resolution (CDR) feature exists. To know more, see Oracle GoldenGate Automatic Conflict Detection and Resolution in the Oracle Database XStream Guide.
9.5 Additional Information
The following documentation provides additional information of relevance to configuring Oracle GoldenGate.
-
For additional system requirements, process configuration, and database setup requirements, see the Oracle GoldenGate installation and configuration document for your database type.
-
For detailed instructions on configuring Oracle GoldenGate change capture and delivery groups, see Configuring Online Change Synchronization.
-
For additional tuning options for Oracle GoldenGate, see Tuning the Performance of Oracle GoldenGate.
-
For complete syntax and descriptions of the Oracle GoldenGate commands and parameters, see Reference for Oracle GoldenGate.
9.6 Creating an Active-Active Configuration
Refer to Figure 9-1 for a visual representation of the objects you will be creating.
Figure 9-1 Oracle GoldenGate Configuration for Active-active Synchronization
Description of "Figure 9-1 Oracle GoldenGate Configuration for Active-active Synchronization"
9.6.1 Prerequisites on Both Systems
Perform these prerequisite tasks on both systems:
-
Create a Replicat checkpoint table (unless using Oracle integrated Replicat). See Creating a Checkpoint Table for instructions.
-
Configure the Manager process. See Configuring Manager and Network Communications for instructions.
Parent topic: Creating an Active-Active Configuration
9.6.2 Configuration from Primary System to Secondary System
These steps add the processes necessary to send data from the primary system to the secondary database.
To Configure the Primary Extract Group
Perform these steps on the primary system.
-
Use the
ADD EXTRACT
command to create a primary Extract group. For documentation purposes, this group is calledext_1
.ADD EXTRACT
ext_1
, {TRANLOG | INTEGRATED TRANLOG}, BEGINtime
-
Use the
ADD EXTTRAIL
command to add a local trail. For documentation purposes, this trail is calledlocal_trail_1
.ADD EXTTRAIL
local_trail_1
, EXTRACText_1
For
EXTRACT
, specify theext_1
group to write to this trail -
Use the
EDIT PARAMS
command to create a parameter file for theext_1
group. Include the following parameters plus any others that apply to your database environment. For possible additional required parameters, see the Oracle GoldenGate installation and setup guide for your database.-- Identify the Extract group: EXTRACT
ext_1
-- Specify database login information as needed for the database: [SOURCEDBdsn_1
][, USERIDALIASalias
] -- Specify the local trail that this Extract writes to -- and the encryption algorithm: ENCRYPTTRAILalgorithm
EXTTRAILlocal_trail_1
-- Exclude Replicat transactions. Uncomment ONE of the following: -- DB2 z/OS, DB2 LUW, DB2 IBM i, and Oracle (classic capture): -- TRANLOGOPTIONS EXCLUDEUSERReplicat_user
-- Oracle (classic capture) alternative to EXCLUDEUSER: -- EXCLUDEUSERIDOracle_uid
-- Oracle integrated capture: -- EXCLUDETAGtag
-- SQL Server: -- TRANLOGOPTIONS EXCLUDETRANStransaction_name
-- -- Teradata: -- SQLEXEC 'SET SESSION OVERRIDE REPLICATION ON;' -- SQLEXEC 'COMMIT;' -- Specify API commands if Teradata: VAM library name, PARAMS ('param
' [, 'param
'] [, ...]) -- Capture before images for conflict resolution: GETBEFORECOLS (ONoperation
{ALL | KEY | KEYINCLUDING (col_list
) | ALLEXCLUDING (col_list
)}) -- Log all scheduling columns for CDR and if using integrated Replicat LOGALLSUPCOLS -- Specify tables to be captured and (optional) columns to fetch: TABLE [container
.|catalog
.]owner
.* [, FETCHCOLScols
| FETCHCOLSEXCEPTcols
];Note:
The VAM parameter in the examples is used only for heterogeneous databases and does not apply to Oracle Database.
To Configure the Data Pump
Perform these steps on the primary system.
-
Use the
ADD EXTRACT
command to create a data pump group. For documentation purposes, this group is calledpump_1
.ADD EXTRACT
pump_1
, EXTTRAILSOURCElocal_trail_1
, BEGINtime
For
EXTTRAILSOURCE
, specifylocal_trail_1
as the data source. -
Use the
ADD RMTTRAIL
command to add a remote trail that will be created on the secondary system. For documentation purposes, this trail is calledremote_trail_1
.ADD RMTTRAIL
remote_trail_1
, EXTRACTpump_1
For
EXTRACT
, specify thepump_1
data pump to write to this trail.See Reference for Oracle GoldenGate for additional
ADD RMTTRAIL
options.Use the
EDIT PARAMS
command to create a parameter file for thepump_1
group. Include the following parameters plus any others that apply to your database environment.-- Identify the data pump group: EXTRACT
pump_1
-- Specify database login information as needed for the database: [SOURCEDBdsn_1
][, USERIDALIASalias
] -- Decrypt the data only if the data pump must process it. -- DECRYPTTRAIL -- Specify the name or IP address of the secondary system -- and optional encryption of data over TCP/IP: RMTHOSTOPTIONSsystem_2
, MGRPORTport_number
, ENCRYPTencryption_options
-- Specify remote trail and encryption algorithm on secondary system: ENCRYPTTRAILalgorithm
RMTTRAILremote_trail_1
-- Specify tables to be captured: TABLE [container
.|catalog
.]owner
.*;
To Configure the Replicat Group
Perform these steps on the secondary system.
-
Use the
ADD REPLICAT
command to create a Replicat group. For documentation purposes, this group is calledrep_1
.ADD REPLICAT
rep_1
[, INTEGRATED | COORDINATED [MAXTHREADSnumber
]] , EXTTRAILremote_trail_1
, BEGINtime
For
EXTTRAIL
, specifyremote_trail_1
as the trail that this Replicat reads. -
Use the
EDIT PARAMS
command to create a parameter file for therep_1
group. Include the following parameters plus any others that apply to your database environment. For possible additional required parameters, see the Oracle GoldenGate installation and setup guide for your database.-- Identify the Replicat group: REPLICAT
rep_1
-- State that source and target definitions are identical: ASSUMETARGETDEFS -- Specify database login information as needed for the database: [TARGETDBdsn_2
][, USERIDALIASalias
] -- Specify error handling rules: REPERROR (error
,response
) -- Set redo tag for Oracle only replicat via settag -- Default is 00. SETTAGtag_value
-- Specify tables for delivery, threads if coordinated Replicat -- and conflict-resolution: MAP [container
.|catalog
.]owner
.*, TARGETowner
.*, COMPARECOLS (ONoperation
{ALL | KEY | KEYINCLUDING (col_list
) | ALLEXCLUDING (col_list
)}), RESOLVECONFLICT (conflict type
(resolution_name
,resolution_type
COLS (col
[,...])) [, THREAD (thread_ID
)] [, THREADRANGE (thread_range
[,column_list
])] ; -- Specify mapping of exceptions to exceptions table: MAP [container
.|catalog
.]owner
.*, TARGETowner
.exceptions
, EXCEPTIONSONLY;
Parent topic: Creating an Active-Active Configuration
9.6.3 Configuration from Secondary System to Primary System
These steps add the processes necessary to send data from the secondary system to the primary database.
To Configure the Primary Extract Group
Perform these steps on the secondary system.
Note:
This is a reverse image of the configuration that you just created.
-
Use the
ADD EXTRACT
command to create a primary Extract group. For documentation purposes, this group is calledext_2
.ADD EXTRACT
ext_2
, {TRANLOG | INTEGRATED TRANLOG}, BEGINtime
-
Use the
ADD EXTTRAIL
command to add a local trail. For documentation purposes, this trail is calledlocal_trail_2
.ADD EXTTRAIL
local_trail_2
, EXTRACText_2
For
EXTRACT
, specify theext_2
group to write to this trail. -
Use the
EDIT PARAMS
command to create a parameter file for theext_2
group. Include the following parameters plus any others that apply to your database environment. For possible additional required parameters, see the Oracle GoldenGate installation and setup guide for your database.-- Identify the Extract group: EXTRACT
ext_2
-- Specify database login information as needed for the database: [SOURCEDBdsn_2
][, USERIDALIASalias
] -- Specify the local trail that this Extract writes to -- and the encryption algorithm: ENCRYPTTRAILalgorithm
EXTTRAILlocal_trail_2
-- Exclude Replicat transactions. Uncomment ONE of the following: -- DB2 z/OS, DB2 LUW, DB2 IBM i, and Oracle: -- TRANLOGOPTIONS EXCLUDEUSERReplicat_user
-- Oracle alternative to EXCLUDEUSER: -- EXCLUDEUSERIDOracle_uid
-- Oracle integrated capture: -- EXCLUDETAGtag
-- SQL Server: -- TRANLOGOPTIONS EXCLUDETRANStransaction_name
---- Teradata: -- SQLEXEC 'SET SESSION OVERRIDE REPLICATION ON;' -- SQLEXEC 'COMMIT;' -- Oracle: -- TRACETABLEtrace_table_name
-- Log all scheduling columns for CDR and if using integrated Replicat LOGALLSUPCOLS -- Capture before images for conflict resolution: GETBEFORECOLS (ONoperation
{ALL | KEY | KEYINCLUDING (col_list
) | ALLEXCLUDING (col_list
)}) -- Specify tables to be captured and (optional) columns to fetch: TABLE [container
.|catalog
.]owner
.* [, FETCHCOLScols
| FETCHCOLSEXCEPTcols
];Note:
To replicate Oracle DBFS data, specify the internally generated local read-write DBFS tables in the
TABLE
statement on each node. For more information on identifying these tables and configuring DBFS for propagation by Oracle GoldenGate, see Applying the Required Patch in Using Oracle GoldenGate for Oracle Database.
To Configure the Data Pump
Perform these steps on the secondary system.
-
Use the
ADD EXTRACT
command to create a data pump group. For documentation purposes, this group is calledpump_2
.ADD EXTRACT
pump_2
, EXTTRAILSOURCElocal_trail_2
, BEGINtime
For
EXTTRAILSOURCE
, specifylocal_trail_2
as the data source. -
Use the
ADD RMTTRAIL
command to add a remote trail that will be created on the primary system. For documentation purposes, this trail is calledremote_trail_2
.ADD RMTTRAIL
remote_trail_2
, EXTRACTpump_2
For
EXTRACT
, specify thepump_2
data pump to write to this trail. -
Use the
EDIT PARAMS
command to create a parameter file for thepump_2
group. Include the following parameters plus any others that apply to your database environment.-- Identify the data pump group: EXTRACT
pump_2
-- Specify database login information as needed for the database: [SOURCEDBdsn_2
][, USERIDALIASalias
] -- Decrypt the data only if the data pump must process it. -- DECRYPTTRAIL -- Specify the name or IP address of the primary system -- and optional encryption of data over TCP/IP: RMTHOSTOPTIONSsystem_1
, MGRPORTport_number
, ENCRYPTencryption_options
-- Specify the remote trail and encryption algorithm on the primary system: ENCRYPTTRAILalgorithm
RMTTRAILremote_trail_2
-- Specify tables to be captured: TABLE [container
.|catalog
.]owner
.*;Note:
To replicate Oracle DBFS data, specify the internally generated local read-write DBFS tables in the
TABLE
statement on each node. For more information on identifying these tables and configuring DBFS for propagation by Oracle GoldenGate, see Configuring the DBFS File System in Using Oracle GoldenGate for Oracle Database.
To Configure the Replicat Group
Perform these steps on the primary system.
-
Use the
ADD REPLICAT
command to create a Replicat group. For documentation purposes, this group is calledrep_2
.ADD REPLICAT
rep_2
[, INTEGRATED | COORDINATED [MAXTHREADSnumber
]] , EXTTRAILremote_trail_2
, BEGINtime
For
EXTTRAIL
, specifyremote_trail_2
as the trail that this Replicat reads.See
ADD RMTTRAIL
in Reference for Oracle GoldenGatefor detailed information about these and other options that may be required for your installation. -
Use the
EDIT PARAMS
command to create a parameter file for therep_2
group. Include the following parameters plus any others that apply to your database environment.-- Identify the Replicat group: REPLICAT
rep_2
-- State that source and target definitions are identical: ASSUMETARGETDEFS -- Specify database login information as needed for the database: [TARGETDBdsn_1
][, USERIDALIASalias
] -- Specify error handling rules: REPERROR (error
,response
) -- Specify tables for delivery, threads if coordinated Replicat -- and conflict-resolution: MAP [container
.|catalog
.]owner
.*, TARGETowner
.*, COMPARECOLS (ONoperation
{ALL | KEY | KEYINCLUDING (col_list
) | ALLEXCLUDING (col_list
)}), RESOLVECONFLICT (conflict type
(resolution_name
,resolution_type
COLS (col
[,...])) [, THREAD (thread_ID
)] [, THREADRANGE (thread_range
[,column_list
])] ; -- Specify mapping of exceptions to exceptions table: MAP [container
.|catalog
.]owner
.*, TARGETowner
.exceptions
, EXCEPTIONSONLY;Note:
To replicate Oracle DBFS data, specify the internally generated local read-write DBFS tables in the
TABLE
statement on each node.
Parent topic: Creating an Active-Active Configuration