Db2 bufferpools and the linux file cache perform largely the same function ; that is , cache a copy of data read from disk Db2缓冲池与linux文件缓存基本上执行相同的功能,也就是缓存从磁盘读出的数据的拷贝。
As noted earlier , the flow of reads under normal circumstances is to read into the file cache and then copy into the bufferpool 如前所述,一般情况下,读操作的流程是先将数据读到文件缓存中,然后再复制到缓冲池。
Direct i o solves this issue by directing reads to be made directly from disk into db2 memory areas , bypassing the file cache entirely 直接i / o直接将数据从磁盘读到db2内存区域,完全绕过文件缓存,从而解决了上述问题。
Prior to the linux 2 . 6 kernel , file caching on a linux system could go awry , using a huge number of cycles trying to manage the file cache 对于linux 2 . 6之前版本的内核, linux系统上的文件缓存可能会出差错,需要花费大量的周期来管理文件缓存。
Memory utilization and file caching are related elements that affect performance and are important to consider when tuning a database system 内存的使用和文件缓存是彼此相关的两个部分,它们一起影响着性能,在对数据库系统进行调优时,要重点考虑这两个方面。