Replication Server

Note: this component is designated "advanced", which means that objects of this type are not expected to be created or altered in most environments. If you believe that such a change is necessary, you may want to contact support in order to understand the potential impact of that change.

Replication Servers publish updates to Directory Server instances within a Replication Domain.

Replication Servers publish updates to Directory Server instances within a Replication Domain. They are created when replication is enabled using the dsreplication tool.

Relations to This Component Properties

Relations to This Component

The following components have a direct composition relation to Replication Servers:

Properties

The properties supported by this managed object are as follows:


Basic Properties: Advanced Properties:
 replication-server-id  je-property
 replication-db-directory  replication-purge-delay
 replication-port  target-database-size
 compression-criteria  listen-on-all-addresses
 gateway-priority  heartbeat-interval
 remote-monitor-update-interval
 restricted-domain
 missing-changes-alert-threshold-percent
 missing-changes-policy
 include-all-remote-servers-state-in-monitor-message

Basic Properties

replication-server-id (Read-Only)

Description
Specifies a unique identifier for the Replication Server. Each Replication Server must have a different server ID.
Default Value
None
Allowed Values
An integer value. Lower limit is 1. Upper limit is 65535 .
Multi-Valued
No
Required
Yes
Admin Action Required
None. Modification requires no further action

replication-db-directory

Description
The path where the Replication Server stores all persistent information.
Default Value
changelogDb
Allowed Values
A filesystem path
Multi-Valued
No
Required
Yes
Admin Action Required
The Directory Server must be restarted for changes to this setting to take effect. Modification requires that the Directory Server be stopped, the database directory manually relocated, and then the Directory Server restarted. While the Directory Server is stopped, the old database directory and its contents must be manually moved or copied to the new location.

replication-port (Read-Only)

Description
The port on which this Replication Server waits for connections from other Replication Servers or Directory Server instances.
Default Value
None
Allowed Values
An integer value. Lower limit is 1. Upper limit is 65535 .
Multi-Valued
No
Required
Yes
Admin Action Required
This modification must also be made to the Server Instance representing this server in the topology registry. This ensures that the most up-to-date information is used by topology-related tools.

compression-criteria

Description
Specifies when the replication traffic should be compressed. The replication traffic can be compressed to reduce the amount of data flowing through particular data links.
Default Value
remote
Allowed Values
always - Replication traffic is always compressed except when the replica and the replication server are in the same process.

remote - Replication traffic compression is enabled when a replica or replication server communicates with a replication server in a different location.

never - Replication traffic is never compressed.
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

gateway-priority

Description
Specifies the gateway priority of the Replication Server in the current location. The Replication Server acting as a gateway handles communication with servers in other locations. The highest priority value is 1 and the lowest value is 10. To prevent this Replication Server from ever assuming the gateway role, set the priority to 0. The gateway role is assumed by the Replication Server with the highest gateway priority. Multiple Replication Servers may, however, share the same priority value. In such case, the first Replication Server requesting the role will normally assume this role. See the Replication chapter in the Administration Guide for further details about the gateway election process.
Default Value
5
Allowed Values
An integer value. Lower limit is 0. Upper limit is 10 .
Multi-Valued
No
Required
Yes
Admin Action Required
None. Modification requires no further action


Advanced Properties

je-property (Advanced Property)

Description
Specifies the database and environment properties for the Berkeley DB Java Edition database for the replication changelog. Any Berkeley DB Java Edition property can be specified using the following form: property-name=property-value. Refer to the Directory Server documentation for further information on related properties, their implications, and range values. The definitive identification of all the property parameters is available in the example.properties file of Berkeley DB Java Edition distribution.
Default Value
None
Allowed Values
A string
Multi-Valued
Yes
Required
No
Admin Action Required
Changes to this configuration property will only take effect if the associated multimaster replication synchronization provider is disabled and re-enabled (or the Directory Server is restarted).

replication-purge-delay (Advanced Property)

Description
Changes are guaranteed to be maintained in the changelog database for at least this duration. Setting target-database-size can allow additional changes to be maintained up to the configured size on disk. The replication purge delay specifies the period after which the Directory Server purges changes in the replication server database. Any change that is older than the purge delay is removed from the replication server database regardless of whether the change has been applied. To ensure proper replication processing, you must have the same purge delay value set for all replication servers in the topology.
Default Value
86400s
Allowed Values
A duration. Lower limit is 0 seconds.
Multi-Valued
No
Required
No
Admin Action Required
Lowering this value will result in the immediate loss of historical replication data. Other servers may enter lockdown mode if they have a larger purge delay and detect missing changes.

target-database-size (Advanced Property)

Description
The replication changelog database is allowed to grow up to this size even if changes are older than the configured replication-purge-delay. Setting this value allows additional changes to be retained in the changelog beyond the purge delay up to the configured total size on disk. It is possible that the changelog database size on disk exceeds this configured value since changes are never purged before the configured replication-purge-delay. To avoid exceeding the configured limit, the server will begin purging data when 95% of the configured limit is reached.

If a backup of the replicationChanges backend is performed as a task, then the size of the database on disk could theoretically exceed this limit since database cleaning cannot occur during the backup. In practice, this will not occur unless replication-purge-delay and target-database-size are configured to very small values and/or the backup is throttled with the --maxMegabytesPerSecond option so that it takes an especially long time to complete.

The server maintains a Gauge that will trigger an Alarm if the size on disk exceeds this limit.

Default Value
Data is purged immediately after the replication-purge-delay expires.
Allowed Values
A positive integer representing a size.
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

listen-on-all-addresses (Advanced Property)

Description
Indicates whether the Replication Server should listen on all addresses for this host. If set to FALSE, then the Replication Server will listen only to the address resolved from the hostname provided.
Default Value
true
Allowed Values
true
false
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

heartbeat-interval (Advanced Property)

Description
Specifies the heartbeat interval that the Directory Server will use when communicating with Replication Servers. The Directory Server expects a regular heartbeat coming from the Replication Server within the specified interval. If a heartbeat is not received within the interval, the Directory Server closes its connection.
Default Value
1m
Allowed Values
A duration. Lower limit is 10000 milliseconds.
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

remote-monitor-update-interval (Advanced Property)

Description
Specifies the duration that topology monitor data will be cached before it is requested again from a remote server. Each replication server has topology monitor data for all other servers participating in replication. This value can be increased for topologies that include a large number servers to reduce the network communication overhead
Default Value
5000ms
Allowed Values
A duration. Lower limit is 1000 milliseconds.
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

restricted-domain (Advanced Property)

Description
Specifies the base DN of domains that are only replicated between server instances that belong to the same replication set. By default, replication changes are propagated in the entire topology. In some cases, however, it is desirable to limit the scope of replication. The entry balancing feature, for example, allows the same base DN to be partitioned into multiple data sets. Changes within a data set may be replicated, while the same is not desired between data sets.
Default Value
None
Allowed Values
A valid DN.
Multi-Valued
Yes
Required
No
Admin Action Required
None. Modification requires no further action

missing-changes-alert-threshold-percent (Advanced Property)

Description
Specifies the missing changes alert threshold as a percentage of the total pending changes. For instance, a value of 80 indicates that the replica is 80% of the way to losing changes. When replication servers connect to each other in order to determine if there are missing changes the latest change on the first server is compared to a range of changes on the second server. If the first server is behind the second server at more than the specified percentage of the way from the newest change on the second server, to the oldest change on the second server, then alert 'replication-missing-changes-risk' is raised.

Setting this property to 0 causes alert 'replication-missing-changes-risk' to be raised if the first server is even one change behind the second server. Setting this property to 100 effectively disables alert 'replication-missing-changes-risk' since exceeding 100% implies there are missing changes, which is handled elsewhere. In most cases this property should be set to at least 20 and at most 90.

If raised consider increasing 'replication-purge-delay' and or 'replication-purge-minimum-retain-count' on all servers so that they have a longer history of changes.

Default Value
80
Allowed Values
An integer value. Lower limit is 0. Upper limit is 100 .
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

missing-changes-policy (Advanced Property)

Description
Determines how the server responds when replication detects that some changes might have been missed. Each missing changes policy is a set of missing changes actions to take for a set of missing changes types. The value configured here acts as a default for all replication domains on this replication server. Each missing changes policy is a set of missing changes actions to take as a function of a set of missing changes types. This makes it possible to select the preferred data integrity to availability tradeoff. Each value offers greater availability than the previous one at the cost of data integrity.
Each missing changes event will result in a 'replication-missing-changes-action' alarm with an associated message that will describe the action taken as a function of the missing changes configuration.
Alternatively 'missing-changes-policy' may be configured on the replication domains if per replication domain configuration is needed.
Default Value
favor-availability
Allowed Values
maximum-integrity - Enter the missing changes lockdown state in response to missing changes as much as possible. This strongly favors data integrity over availability.

favor-integrity - This is the same as the 'maximum-integrity' policy except there is increased availability with regard to 'indirect' missing changes (the replica is hosted by a server not involved in the connection). Enter the missing changes lockdown state in those case where it makes sense to do so.

favor-availability - This is the same as the 'favor-integrity' policy except there is increased availability with regard to 'previous' missing changes (also known as 'null CN' missing changes). With this policy the missing changes state will not persist across restarts of the server. This is the recommended balance between data integrity and availability for most installations.

maximum-availability - Never enter the missing changes lockdown state in response to missing changes. This strongly favors availability over data integrity. Note that this value allows changes to be applied from remote servers even though some of those changes are missing, which will cause the replication servers to enter a persistently inconsistent state. To restore the missing changes initialize from a known good server with 'dsreplication initialize'. Alternatively the missing changes can be determined by comparing to a known good server with 'ldap-diff' and then apply the missing changes using the replication repair control.
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action

include-all-remote-servers-state-in-monitor-message (Advanced Property)

Description
Indicates monitor messages should include information about remote servers. Replication server communication includes monitor messages that describe the state of the sending server. When Replication Server is set monitor messages will also include information about remote (other) servers in the topology. Normally information about remote servers is not needed since each server describes itself, but including it allows greater tolerance for lost monitor messages.

Setting Replication Server to 'false' may be particularly helpful in large topologies since otherwise the size of monitor messages scales with the number of servers.

Default Value
false
Allowed Values
true
false
Multi-Valued
No
Required
No
Admin Action Required
None. Modification requires no further action