SharePlex 9.2.8 - Installation and Setup for Oracle Source

Routing specifications in a configuration file

The following instructions show you how to build a routing map based on where you want to send the source data. A routing map sends replicated data to the correct target on the correct target system, or systems.

For details about the components of these configurations, see:

Database specifications in a configuration file

Target specifications in a configuration file

Routing to one target

A simple routing map sends replicated data from one source object to one target object.

Routing to a cloud service

There are special routing requirements for database targets that are hosted by a cloud service such as EC2 and RDS on Amazon AWS, Azure SQL in Microsoft Azure, and Compute Virtual Machines in Oracle Cloud Infrastructure. Whether the service is Infrastructure as a Service (IaaS) or Platform as a Service (PaaS) makes a difference in how you install and configure SharePlex. The following explains these requirements.

IaaS targets

If replicating to a database target hosted in an IaaS cloud service, specify the full endpoint URL as the target host in the routing map.

For example, the following routing map routes to a cloud database on Amazon EC2:

ec2-12-345-678-910.compute-1.amazonaws.com@o.myora

Alternately, you can map the private IP address of the cloud service to a short name in the local hosts file, and then specify that name as the host in the routing map, for example:

shortname@o.myora

PaaS targets

If replicating to a database target hosted in a PaaS cloud service, there are special installation, setup, and routing requirements. Because SharePlex cannot be installed directly on a PaaS cloud server, you must install SharePlex on either the source server or an intermediary server, from which Post connects to the target cloud database. For more information, see Installation and setup for cloud-hosted databases in the SharePlex Installation and Setup Guide.

Routing to multiple targets

A compound routing map sends replicated data from one source object to multiple target objects. It enables you to specify the source and target objects once for all routes, rather than type a separate configuration entry for each route. Only one target specification can be used in a compound routing map, so all of the target objects must be identical as follows:

• All are of one type: All the same database object type or all a JMS queue or all a JMS topic or all a Kafka topic, or all a file (but no combination of these).
• All have the same fully qualified name, including any table specifications in a JMS, Kafka, or file target specification.

• All have identical column mappings or key mappings, if used. For more information about these mappings, see:

  • Define a unique key
  • Map source and target columns

NoteS:

Routing between objects on the same system

You can replicate between the following:

• For Oracle, you can replicate between objects that are in the same database or in different databases on the same system. You can replicate between objects that have the same name, so long as their owners are different.

When SharePlex replicates between objects on the same system, it does not create Import and Export processes. You can force SharePlex to create Import and Export processes by using the following routing map. If you do not need the Import or Export processes, omit the host* portion of the routing map.

Configuration with replication to objects in the same or different database on the same system

Routing Limitations

• By default, SharePlex supports replication to a maximum of 19 direct target systems. That is the maximum number of processes that can read the export queue. To replicate to more than 19 targets, use named export queues. With each additional queue that you add, you can replicate to 19 additional targets. For more information, see Configure named export queues.
• Each instance of sp_cop on a system permits a maximum of 1024 different routes. This limitation includes each route that uses a different named post queue (see Configure named post queues.) If your replication strategy requires more than 1024 routes, consider using one or more intermediary systems to divide the routes among multiple sp_cop instances. For more information, see Configure replication to share or distribute data.
• By default, each sp_cop instance allows a total of 25 queues on a system. There will always be one capture queue on a source system and one post queue on a target. Therefore, you can have as many as 24 named export queues on a source system and 24 named post queues on a target system. If a system serves as both a source and target, you will have both a capture queue and a post queue. That allows you to create up to 23 named queues of either type (or a mix of both). If system memory permits, you can change the number of allowed queues by setting the SP_QUE_MAX_QUEUES parameter. See the SharePlex Reference Guide for more information about this parameter.