db file sequential read

1.获取最大的磁盘读的ＳＱＬ语句

SELECT *
  FROM (SELECT PARSING_USER_ID EXECUTIONS,
               SORTS,
               COMMAND_TYPE,
               DISK_READS,
               sql_text
          FROM v$sqlarea
         ORDER BY disk_reads DESC)
 WHERE ROWNUM < 20;

2.关注v$session_wait视图

对于db file sequential read中,p1指数据文件ID,p2指block#号,p3指读取的block数量,这个事件一般不可避免，大多由于ＳＱＬ使用索引不当，造成从磁盘上读取连续的数据，接近于全表扫描.可以通过db_file_multiblock_read_count参数来调整每次读取的block数,减少IO.