Chinaunix首页 | 论坛 | 博客
  • 博客访问: 1611518
  • 博文数量: 292
  • 博客积分: 10791
  • 博客等级: 上将
  • 技术积分: 2479
  • 用 户 组: 普通用户
  • 注册时间: 2010-03-20 21:06
文章分类

全部博文(292)

文章存档

2011年(31)

2010年(261)

分类:

2011-02-16 13:46:33

bread/s, bwrit/s【物理读】
    Reports the number of block I/O operations. These I/Os are generally performed by the kernel to manage the block buffer cache area, as discussed in the description of the lread/s value.
lread/s, lwrit/s【逻辑读】
    Reports the number of logical I/O requests. When a logical read or write to a block device is performed,a logical transfer size of less than a full block size may be requested. The system accesses the physical device units of complete blocks and buffers these blocks in the kernel buffers that have been set aside for this purpose (the block I/O cache area). This cache area is managed by the kernel, so that ultiple logical reads and writes to the block device can access previously buffered data from the cache  and require no real I/O to the device. Application read and write requests to the block device are reported statistically as logical reads and writes. The block I/O performed by the kernel to the block device in management of the cache area is reported as block reads and block writes.

逻辑读从缓冲中读,如果缓冲中没有数据则发生物理读。

所以缓冲命中率为%rcache:[(100)x(lreads -  breads)/ (lreads)].
阅读(2760) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~