ogg 抽取进程是在8点停止抽取（比如当时的SCN是 100），在容灾切换之后依然需要从8点的（SCN 100 对应）归档日志抽取数据才能保证数据不丢失

SCN不连续，在原主库上想辙,看有无可能apply到DG切换点前，估计是个坑

8点钟 手动stop了抽取进程，这是记录查看info 抽取进程，查看时间点，切换后抽取也应该是该时间点。
你上面的报错是缺少归档日志文件了？
请尝试做下日志切换，然后再抽取。
参考OGG-00685 Begin Time Prior to 1) This is a startup bug error where certain initial log positions are not resolved at initialization.
This should happen only if the first log switching didn’t occur after the db was created and before extract was started

It is bug (11839680).
2) There are THREADs begin added or deleted.
SOLUTION
It is resolved by performing a database log switch.
SELECT MIN(TO_CHAR (first_time, ‘YYYY-MON-DD HH24:MI:SS’)) FROM v$log_history WHERE thread# = <for each thread>; can be done to determinmg the minumum time.
Other causes and solutions:
If some of the nodes in RAC are not active, OGG will hit this error when trying to find the redo logs for those inactive nodes.
The solution in this case is to exclude those inactive nodes from GG, e.g. THREADOPTIONS PROCESSTHREADS SELECT 1
REFERENCES