Timestamp Event Replication

The driver implements the replication of event data between redundant driver instances in order to be certain that all TS events are pushed to the Alarm and Tran systems.

Event replication helps in a case when there is a lack of communication between the I/O server which receives events from the OFS and the Alarm Server.

For example in a system which has primary I/O server, standby I/O server and alarm server, the primary I/O server receives events from the OFS server. It pushes them to the alarm server and also replicates them to the standby I/O server. The standby I/O server stores these events in the event buffer.

In the case of a lack of communication between the primary I/O server and the alarm server, the primary I/O server will not be able to push events to the alarm server.

If during that time switchover happens (for example the primary I/O server was shut down), the standby I/O server will push its event buffer to the alarm server. The standby I/O server will store events data for 10 minutes. Also the size of the event buffer is limited to 300,000 events.

In a normal scenario when there is no lack of communication, the primary I/O server sends a confirmation message to the standby server after an event has been successfully propagated through the alarm system. On receiving that message, the standby I/O server deletes the corresponding event from the event buffer.

Also, if there is a lack of communication between the primary and standby I/O servers and the replication of events is not possible, the primary I/O server stores events received from the OFS for a limited time (10 min) and sends them to the standby I/O server after connection has been restored.