Scheduling synchronization

You can set up dbmlsync to synchronize periodically based on rules you define. There are two ways you can set this up:

Use the dbmlsync extended option SCHEDULE to initiate synchronization at specific times of the day or week or at regular intervals. In this case, dbmlsync remains running until stopped by the user.

See Setting up scheduling with dbmlsync options.
Use dbmlsync event hooks to initiate synchronization based on logic that you define. This is the best way to implement synchronization at irregular intervals or in response to an event. In this case, you can stop dbmlsync programmatically from your hook code.

See Initiating synchronization with event hooks.

This method is not available when the Dbmlsync API or the SQL SYNCHRONIZE statement is used.

Hovering

When hovering, dbmlsync scans the database transaction log and builds its upload during the delay between synchronizations. This allows synchronization to proceed more quickly when it is triggered because some of the work is already done.

When hovering, dbmlsync scans to the end of the transaction log then polls the log periodically for new transactions. You can control the interval between polls using the PollingPeriod extended option of the -pp option. See PollingPeriod (pp) extended option.

When you are hovering on two or more subscriptions at the same time, you can use the HoverRescanThreshold extended option or the sp_hook_dbmlsync_log_rescan event hook to limit memory usage by recovering otherwise lost memory.

Hovering can be disabled using the DisablePolling extended option or the -p option.

For more information, see:

Setting up scheduling with dbmlsync options
Initiating synchronization with event hooks

Discuss this page in DocCommentXchange.