我使用struts2创建了一个web应用程序,使用mysql创建了hibernate。我迁移到播放框架,所以现在我有两个应用程序实例,其中一个用struts2编写。我可以看到游戏的性能大幅提升,但是当我使用JMeter来模拟加载时,struts2版本的缩放比使用更低的延迟更好。为什么这样,这是否意味着坚持使用struts2更好? webapp(我从localhost测试,两个应用程序都有相同的代码,除了模板语言,struts2在tomcat7上运行,而在独立服务器上运行时运行)
Struts 2 Version
Start time Tread Name Label time(ms) Status Bytes Latency
1 10:24:49.854 videovixx struts2 1-6 HTTP Request 981 Success 9116 6
2 10:24:49.830 videovixx struts2 1-5 HTTP Request 1101 Success 9116 16
3 10:24:49.780 videovixx struts2 1-3 HTTP Request 1264 Success 9116 2
4 10:24:49.774 videovixx struts2 1-2 HTTP Request 1309 Success 9116 15
5 10:24:49.741 videovixx struts2 1-1 HTTP Request 1509 Success 9116 10
6 10:24:49.868 videovixx struts2 1-7 HTTP Request 1513 Success 9116 43
7 10:24:49.933 videovixx struts2 1-10 HTTP Request 1471 Success 9116 5
8 10:24:50.294 videovixx struts2 1-27 HTTP Request 1156 Success 9116 134
9 10:24:50.037 videovixx struts2 1-15 HTTP Request 1503 Success 9116 68
10 10:24:49.950 videovixx struts2 1-11 HTTP Request 1608 Success 9116 18
Play Version
Start time Tread Name Label time(ms) Status Bytes Latency
1 10:24:49.809 videovixx play 2-1 HTTP Request 881 Success 13016 881
2 10:24:49.803 videovixx play 2-2 HTTP Request 1562 Success 13016 1562
3 10:24:49.839 videovixx play 2-3 HTTP Request 2146 Success 13016 2146
4 10:24:49.920 videovixx play 2-7 HTTP Request 2065 Success 13016 2065
5 10:24:50.080 videovixx play 2-15 HTTP Request 1905 Success 13016 1905
6 10:24:50.242 videovixx play 2-23 HTTP Request 1743 Success 13016 1743
7 10:24:50.493 videovixx play 2-36 HTTP Request 1492 Success 13016 1492
8 10:24:50.570 videovixx play 2-39 HTTP Request 1619 Success 13016 1619
9 10:24:50.646 videovixx play 2-43 HTTP Request 1571 Success 13016 1571
13016 2105
答案 0 :(得分:0)
您需要根据应用程序的配置文件配置线程模型。更具体地说,您的应用程documentation建议采用以下配置:
play {
akka {
event-handlers = ["akka.event.slf4j.Slf4jEventHandler"]
loglevel = WARNING
actor {
default-dispatcher = {
fork-join-executor {
parallelism-min = 300
parallelism-max = 300
}
}
}
}
}