MySQL索引差异

Question

我有一个查询在两个不同的mysql服务器实例中有不同的性能，两者都有明显相同的表和索引。这是查询：

select t.id, t.artist, t.title, count(*) from `raw_detection` r, tracks t where t.id = r.`track_id`
and r.duplicated = false
and (t.title = '' or t.title is null or t.artist = '' or t.artist is null)
group by 1,2,3
order by count(*) desc;

当我使用此查询执行explain时，快速服务器会响应：

id  s_type  table   type  possible_keys    key  key_len ref rows Extra   
1   SIMPLE  t   index_merge PRIMARY,index_tracks_on_title,index_tracks_on_artist    index_tracks_on_title,index_tracks_on_artist    602,602 NULL    11812   Using sort_union(index_tracks_on_title,index_tracks_on_artist); Using where; Using temporary; Using filesort
1   SIMPLE  r   ref index_raw_detection_on_duplicated_and_created_at,index_raw_detection_on_track_id    index_raw_detection_on_track_id 4   playax.t.id 84  Using where

慢速服务器不同：

id  s_type  table   type  possible_keys    key  key_len ref rows Extra   
1   SIMPLE  t   ALL PRIMARY,index_tracks_on_title,index_tracks_on_artist    NULL    NULL    NULL    177313  Using where; Using temporary; Using filesort
1   SIMPLE  r   ref index_raw_detection_on_duplicated_and_created_at,index_raw_detection_on_track_id    index_raw_detection_on_track_id 4   playax.t.id 20  Using where

为什么会这样？