|
|
|
|
|
When you take a snapshot on the Secondary, it contains a point-in-time copy of the data on the Primary. Because the Secondary may be behind the Primary, it is not known exactly what time on the Primary this point-in-time copy represents.
VVR maintains a block-level consistency between the Primary and Secondary data volumes. But applications, for example, a file system, that use the data volumes have a higher-level consistency requirement. To support this, VVR provides the IBC facility.
IBC messaging enables you to send a message in the replication stream to notify the Secondary that an event has occurred on the Primary. In the case of a file system, you can use the sync command on the Primary, and then send an IBC message. When this message arrives on the Secondary, the data on the Secondary is consistent at the file system level and replication stops. Therefore, further updates are not applied to Secondary data volumes but are stored in the Secondary SRL. You then split off a mirror, which now contains a consistent image of the file system, and unfreeze replication. After the unfreeze operation all the pending updates that are stored in the secondary SRL are applied to the secondary data volumes.
The model with IBC Messaging is that a process on the Secondary waits for the IBC Message and a process on the Primary sends the message when the desired event has occurred.
Note
If you choose not to use IBC Messaging, the data on the Secondary is consistent and can be recovered by the application but it might be out-of-date and potentially out of sync.