The snapshot concept was introduced from the 11g Version, which allows the use of a physical standby database in the read-write mode for a short period of time. You can convert a physical standby to a snapshot standby by using either traditional SQL *Plus or using the Data Guard broker and grid control at any time. Even if you convert it to a snapshot standby, it will still receive data continuously from the production database archive, so that in the next conversion from a snapshot to a physical standby it will be used for recovery. In case you have performed recovery at any point in time, the new incarnation will be started. Even though a new incarnation has started, the snapshot standby database will still continue accepting redo from the primary database. If your standby database is clustered and has more than one node, then shut down all the auxiliary RAC instances of the standby prior to performing a snapshot. Note that you should not put a standby database in the snapshot mode for a long time; it results in huge archive logs between the production database and the standby database and in case of critical databases it can have a serious impact.