SharePlex 9.0 - Administration Guide

Create and populate SharePlex groups on Unix and Linux

Where and when to create the SharePlex groups on Unix and Linux depends on whether you install SharePlex as a root or non-root user.

• If you install as non-root, create the groups in the /etc/group file before you run the SharePlex installer. In a cluster, create them on all nodes.*
• If you install SharePlex as a root user, you can direct the installer to create the groups in the /etc/group file. If you install in a cluster, the installer creates the groups on the primary node, but you must create them yourself on the other nodes.

* The groups must exist because the installer adds the SharePlex Administrator user to the spadmin group during the installation process. In a cluster, this user is only added to the primary node. You must add the SharePlex Administrator user to the other nodes.

To create the groups in /etc/group

To assign a user to a group

1. Open the /etc/group file.

2. Add the Unix or Linux user name to the appropriate group. To assign a list of user names to a group, use a comma-separated list (see the following example).

   spadmin:*:102:spadmin,root,jim,jane,joyce,jerry

   If the password field is null, no password is associated with the group. In the example, the asterisk (*) represents the password, “102” represents the numerical group ID, and spadmin is the group. The group ID must be unique.

3. Save the file.

Users can verify their authorization levels by issuing the authlevel command in sp_ctrl.

Create and populate SharePlex groups on Windows

On Windows, the SharePlex groups are created in the Windows User Accounts control panel by the SharePlex installer. To assign users to these groups, use that control panel after you install SharePlex.

Users can verify their authorization levels by issuing the authlevel command in sp_ctrl.

Activate replication in your production environment

This chapter contains instructions for the initial startup of replication from a source database to a target database on production systems.

Contents

What is activation?

When you activate a configuration, through the activate config command in sp_ctrl, SharePlex does the following:

• Activates (read) the configuration file to build a series of internal structures that identify objects and routes. Only one configuration can be active for any given datasource at a time. Configurations for different datasources on a system can be active at the same time. For example, you can activate a configuration for each Oracle instance or SQL Server database on a system.
• Starts the processes that maintain the capture and replication of source transactions.

• (Oracle only) Reconciles replicated data with the copied data. SharePlex applies transactions that occurred after the copy was taken and discards transactions that occurred before the copy (and thus were applied by the copy and would be redundant if applied by replication). The reconcile process is only required for procedures that start replication while the source database is active.

  Note: A SQL Server source database must be quiesced (no transaction activity allowed) while a configuration is being activated.

The activation of a configuration generally proceeds as follows.

1. Assign an activation ID

SharePlex assigns an activation ID number to each configuration activation and its associated replication processes and queues. A configuration can be activated many times, and this ID keeps track of each one.

2. Create an object cache

SharePlex builds an object cache that records the standard metadata needed to support replication: the name, size, and type of columns, NOT-NULL constraints, and whether a column is part of a key. For tables using partitioned replication, additional information is stored.

3. Add a configuration change marker

SharePlex places a configuration-change marker in the data stream. This marker directs sp_cop to generate a new set of replication processes and queues. If another configuration is active for the same datasource, the marker deactivates it, causing the removal of the old processes and queues after the data they contain is posted.

4. Lock the tables, add the activation marker, unlock

(Oracle only) SharePlex locks the tables that are listed in the configuration file so that it can obtain information about them while they are in a read-consistent state. As many tables can be locked concurrently as there are locking threads available. When SharePlex locks a table, it places an activation marker in the data stream that tells the Capture process to start (or stop) replicating that table.

Note: If an application uses NOWAIT locking on tables in the replication configuration, the NOWAIT could fail if it attempts to obtain a lock on an object that is already locked because it is being activated.

SharePlex locks the following:

• All tables added to replication (new and reactivated configurations)
• All tables removed from replication (reactivated configurations)
• All tables where routes changed (reactivated configurations)

Each table is locked for a very short time, just long enough to activate a table. Replication of each table begins as soon as its activation is complete. Should one or more table fail to activate, SharePlex continues with the activation of the other tables. Users can access the data in a source table when the activation lock is released.