Chinaunix首页 | 论坛 | 博客
  • 博客访问: 1740126
  • 博文数量: 310
  • 博客积分: 6853
  • 博客等级: 准将
  • 技术积分: 2833
  • 用 户 组: 普通用户
  • 注册时间: 2005-08-04 16:41
文章分类

全部博文(310)

文章存档

2013年(17)

2012年(42)

2011年(76)

2010年(71)

2009年(99)

2007年(2)

2006年(1)

2005年(2)

分类: 系统运维

2012-06-01 10:06:38

> with mod_jk 1.2.23 status worker there is the output > > Next Maintenance > > The first seems to be the countdown until next maintenance. But what does the second number stand for? Could not find it in the docs.

Due to the way the maintenance feature is implemented, we can't be sure, when *exactly* global maintenance it is happening. The two values are the earliest possible and the latest possible time it will take for the next global maintenance. Global maintenance is responsible for error recovery (allowing workers in error state to be used again for requests) and decaying load values for the balancer. The min and max times will only be correct, if there are any requests at all during that time. The uncertainty comes from the fact, that there is local and global maintenance. Local means once *per web server process* (closing of idle backend connections), global means once per web server. Mladen added a separate maintenance thread for the next version 1.2.27 (off by default), that would make timing much more predictable.
阅读(695) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~
评论热议
请登录后评论。

登录 注册