The default value is 0, which means that the LGWR will wait up to the number of seconds specified by the NET_TIMEOUT attribute of the LOG_ARCHIVE_DEST_n parameter for each SYNC standby destination.

For example, in an Oracle Data Guard configuration that has two standbys which receive redo in SYNC mode, you would define LOG_ARCHIVE_DEST_2 and LOG_ARCHIVE_DEST_3 to use SYNC transport with DATA_GUARD_SYNC_LATENCY set to a value of 2 seconds. When redo needs to be sent to the standbys, it is sent to both the LOG_ARCHIVE_DEST_2 and LOG_ARCHIVE_DEST_3 connections in parallel and the LGWR will wait for acknowledgement. If LOG_ARCHIVE_DEST_2 responds with its message acknowledging that the redo has been received first, the LGWR will not wait for LOG_ARCHIVE_DEST_3 to respond for more than 2 extra seconds (DATA_GUARD_SYNC_LATENCY). If LOG_ARCHIVE_DEST_3 does not respond within those 2 seconds the LGWR disconnects from LOG_ARCHIVE_DEST_3, putting it into the error state. Redo continues to be sent to LOG_ARCHIVE_DEST_2 as usual. After the duration of the REOPEN attribute for LOG_ARCHIVE_DEST_3 has elapsed, the LGWR reconnects to LOG_ARCHIVE_DEST_3 and redo transfer continues. In this manner, the maximum impact of having the second SYNC standby would be restricted to 2 seconds more than having only one SYNC standby, while allowing the Oracle Data Guard configuration to maintain the desired protection level.

LGWR will never wait longer than the value of the NET_TIMEOUT attribute of the LOG_ARCHIVE_DEST_n parameter, regardless of the value of DATA_GUARD_SYNC_LATENCY.