Replication Overview
Replication provides a method for copying data from one location to another. Replication applies to the services themselves, not individual data records. The advantages of replication are two-fold:
- Replication provides a backup; and
- Replication can be used to centralize data.
Centralizing data can be especially beneficial when the data would otherwise have to be accessed from remote sites that have low-bandwidth, high-latency data links such as satellites or dial-up modems.
Keep in mind that service replication processes results in increased network traffic and system load. See Service Replication for more information.
What Data is Replicated?
Replication provides a method for copying data from one location to another. When replication is enabled, any new value (whether the value is the result of an edit, addition, or delete) is written to a change queue by the source service, which is stored in memory. Our replication model includes replication of all service data including values written to each service's change queue.
Delayed CAS alarms and GNS notifications records are replicated from a source site to a destination site. In the event of a failover these records will be available on the standby site after the failover is complete.
CygNet also provides a ReplValidator utility to help you compare records between source and replicated services. See ReplValidator Utility for more information.
Replication Terminology
The following terms describe the replication process for CygNet Software.
| Item | Description |
|---|---|
|
Change queue |
When replication is enabled, any new value (whether the value is the result of an edit, addition, or delete) is written to a change queue by the source service. The change queue is stored in memory. Each item in the change queue is given a unique sequence number (ID) by the source service. The CHANGE_QUEUE_SIZE configuration keyword specifies the size of the change queue. The CHANGE_QUEUE_NEWEST info item represents the sequence number of the last change on the change queue. The info item will return a blank if there are no entries on the change queue. This behavior matches the behavior of the value returned when asking for the last replicated sequence number, represented by the REPL_LAST_SEQ info item, which also returns blank if there are no change queue entries. These info items are used to determine the need to perform a full sync after a failover. |
|
Change queue mapping |
Each replicated service creates and stores a bidirectional change queue map, which is used to translate the sequence IDs in its own change queue with those in the change queue of its source. The last sequence number used in an incremental replication cycle is always mapped and is represented by REPL_LAST_SEQ. The change queue maps are themselves replicated and stored alongside the change queue map of the replicated service. The mapping is persisted on shutdown and restored on startup. |
|
Data sync |
Upon start up, a replicated service performs a full data synchronization with its source. All data in the service is replicated. |
|
Destination service |
An unlimited number of replicated service destinations can be configured to pull data from the source service. Contact your Account Manager or CygNet Sales for licensing information. |
|
Read/Write status |
Replicated services are always read-only. |
|
Replicated service |
The destination service pulling data from the source service. |
|
Replicated service name and Domain ID |
The replicated service pulling the data must have the same name as the source service but must reside on a different host server, running under a different Domain ID. |
|
Replicating service |
The source service from where the data is pulled. |
|
Replication method |
A replicated service pulls data from the source service. |
|
Resynchronization (ReSync) |
If a replicated service or its source service is shut down, or the change queue rolls over, a full resynchronization of all records is performed, with one exception: the VHS never performs a full resynchronization. |
|
Source service |
The service from where replicated data is pulled. The REPL_SOURCE keyword specifies the name of the source service. Note: For replication to function correctly, you must configure the Preferred ARS for the domain ID of the source service where replicated data is being pulled. The Preferred ARS is configured in the CygNet Domain Connection utility. |
More:
