LGWR SYNC步骤

LGWR SYNC redo transport is a synchronous process that causes the database to delay acknowledgment of user commits until the following actions have been completed:

1. LGWR writes redo to the online redo log at the production database (when LGWR SYNC is not used, user commits are acknowledged once this step completes except when the parameter COMMIT NOWAIT is used, as described in Appendix C).
2. The Data Guard LNS process on the production database performs a network send to the Data Guard RFS process on the standby database. For redo write sizes larger than 1MB LNS will issue multiple network sends to the RFS process on the standby.
3. The RFS process receives the redo being sent by LNS and completes the I/O into the standby redo log.
4. RFS process sends acknowledgment back to LNS that the redo has been received and written to disk.
5. LNS posts LGWR that the all the redo has been successfully received and written to disk by the standby.
   On the production database the time for step one is represented with the “log file parallel write” wait event. Steps 2 through 5 are represented with the “LNS wait on SENDREQ” wait event. You can further divide the “LNS wait on SENDREQ” wait event into network time and RFS I/O time by subtracting the “RFS write” wait event obtained on the standby. These wait events can be assessed on the standby by using the query described in the “Data Guard Specific Wait Events” section above for a physical standby or by using AWR for a logical standby.

From: https://community.oracle.com/thread/1053657