Partitions are logical divisions of the Novell® eDirectory database that form a distinct unit of data in the eDirectory tree for administrators to store and replicate eDirectory information. Each partition consists of a Container object, all objects contained in it, and information about those objects. Partitions do not include any information about the file system or the directories and files contained there.
Instead of storing a copy of the entire eDirectory database on each server, you can make a copy of the eDirectory partition and store it on many servers across the network. Each copy of the partition is known as a replica.You can create any number of replicas for each eDirectory partition and store them on any server. There are different types of replicas:
These contain all objects and attributes for a particular partition.
These are used for tree connectivity.
Filtered replicas contain a subset of information from the entire partition, consisting of only the desired classes and attributes which are defined by the server’s replication filter. This filter is used to identify the classes and attributes allowed to pass during inbound synchronization and local changes.
Filtered replicas allow administrators to create sparse and fractional replicas.
The functionality of filtered replicas enables fast response when the data stored in eDirectory is procured by applications. Filtered replicas also allow more replicas to be stored on a single server.
These replicas allow local modifications to classes and attributes that are a subset of the server’s replication filter. However, these replicas can create objects only if all mandatory attributes for the class are within the replication filter.
These replicas do not allow local modifications.
For more information, see Managing Partitions and Replicas in the eDirectory Administration Guide.
When you create partitions, you make logical divisions of your tree. These divisions can be replicated and distributed among different eDirectory servers in your network. When you create a new partition, you split the parent partition and then have two partitions. The new partition becomes a child partition. For example, if you choose an Organizational Unit and create it as a new partition, you split the Organizational Unit and all of its subordinate objects from its parent partition.
The Organizational Unit you choose becomes the root of a new partition. The replicas of the new partition exist on the same servers as the replicas of the parent, and objects in the new partition belong to the new partition’s root object.
In Roles and Tasks, select > .
Type the name and context of the container or use the search feature to find it, and click .
After a few seconds, a confirmation message appears: “Complete: Create Partition.”
When you merge a partition with its parent partition, the chosen partition and its replicas combine with the parent partition. You do not delete partitions; you only merge and create partitions to define how the directory tree is split into logical divisions.
There are several reasons to merge a partition with its parent:
Before merging a partition, check the partition synchronization of both partitions and fix any errors before proceeding. By fixing the errors, you can isolate problems in the directory and avoid propagating the errors or creating new ones. Make sure that all servers which have replicas (including subordinate references) of the partition you want to merge, are up before attempting to merge a partition. If a server is down, eDirectory won’t be able to read the server’s replicas and won’t be able to complete the operation.
If you receive errors in the process of merging a partition, resolve the errors as they appear. Don’t try to fix the error by continuing to perform operations; doing so only results in more errors.
To merge a child partition with its parent partition:
In Roles and Tasks, select > .
Type the name and context of the partition or use the search feature to find it, and click .
After a few seconds, a confirmation message appears: “Complete: Merge Partition.”
Moving a partition lets you move a subtree in your directory tree. You can move a partition root object (which is a Container object) only if it has no subordinate partitions.
When you move a partition, you must follow eDirectory containment rules. For example, you cannot move an Organizational Unit directly under the root of the current tree, because the root’s containment rules allow Locality, Country, or Organization, but not Organizational Unit.
When you move a partition, eDirectory changes all references to the partition root object. Although the object’s common name remains unchanged, the complete name of the container (and of all its subordinates) changes.
In Roles and Tasks, select > .
Type the name and context of the object or use the search feature to find it.
Type the location you want to move the partition to in the Move to text box.
Select the check box and click .
This allows any operations that are dependent on the old location to continue uninterrupted until you can update those operations to reflect the new location. Users can continue to log in to the network and find objects in the original directory location.
The Alias object that is created has the same common name as the moved container and references the new complete name of the moved container.
Make sure your directory tree is synchronizing correctly before you move a partition. If you have any errors in synchronization in either the partition you want to move or the destination partition, do not perform a move partition operation. First, fix the synchronization errors. After moving the partition, if you don’t want the partition to remain a partition, merge it with its parent partition.
Viewing information about a replica tells you about its state. An eDirectory replica can be in various states depending on the partition or replication operations it is undergoing.
To view information about a replica:
In Roles and Tasks, select > .
Select the partition or server whose replica table you want to view, and click .
A table appears listing the replica Partition, Type, Filter, and State. The following list describes the replica states that you might see in iManager:
For more information, see Replica View in the eDirectory Administration Guide.
Viewing information about a partition (from the eDirectory Administration Guide) displays its synchronization information: last successful synchronization and last attempted synchronization.
In Roles and Tasks, select > .
Type the name and context of the partition object, or use the search feature to find it, and click .
Administrators generally use the filtered replica capability to create an eDirectory server that holds a set of filtered replicas that contain only specific objects and attributes that they want synchronized. Filtered replicas maintain a filtered subset of information from an eDirectory partition (objects or object classes along with a filtered set of attributes and values for those objects).
The filtered replica wizard steps you through the configuration of the filtered replicas on the selected server.
In Roles and Tasks, select > .
Type the name and context of the server on which you want to configure a filtered replica, or use the search feature to find it, and click .
Click to define the classes and attributes for a filter set on the selected server.
The replication filter contains the set of eDirectory classes and attributes you want to host on this server’s set of filtered replicas.
Click , then click .
For more information, see Setting Up and Managing Filtered Replicas in the eDirectory Administration Guide.