oracle为何选用hash bucket维护buffer block

Yong Huang

myownstars: Thanks for the two links. The first link talks about modified LRU lists and how new blocks are inserted and old blocks age out. But there's no mention of searching the lists. Is it a simple linear search? Or it uses some more sophisticated data structure?

The second link talks about adaptive hash index, which is an index automatically built (and dropped?) *on an existing index* of a table. If the table does not have any index, there will never be an adaptive hash index. Do I understand that correctly? I think in this discussion here, we should focus on finding the data structures and algorthms that are used to locate data blocks in buffer pool, not on MySQL's indexes. Right?

myownstars

Yong Huang 发表于 2014-5-1 22:41
myownstars: Thanks for the two links. The first link talks about modified LRU lists and how new bloc ...

>we should focus on finding the data structures and algorthms that are used to locate data blocks in buffer pool
正解，但是google很长时间没有发现类似讨论，而且关于Innodb的资料本来就很少，可能需要阅读源码解惑了(个人尝试过但是没有读懂)