线上代码压测问题梳理排查总结.docxVIP

序近期在对阿里云服务器做压力测试,因为webbench ,ab两个工具的压力测试结果和loadrunner,jemer的压测结果相关太远,有上百倍的差距,也是让我们百思不得其解,非常干扰思路.所以对ab,webbench做了简单测试~压测工具简介:工具简介级别Abab - Apache HTTP server benchmarking tool轻量级(比webbench稍强大)Webbenchwebbench - simple forking web benchmark轻量级jmeterApache JMeter是Apache组织开发的基于Java的压力测试工具中级loadrunnerHP研发,收费版,但国内早破解泛滥,有window(6G),linux(1.6G)版本重量级/专业级------loadrunner几尽为国内测试人员的专用压测工具,他的高度灵活性,数据分析功能,图表展示,用户行为模拟等功能非常强大,专业压测还是以loadrunner为准------并发1000概念理解这里需再强调下并发的概念,以并发1000为例:这里的并发1000是针对用户来说,而非服务器每秒并发请求数.即每秒有1000用户(每秒有多次或多种行为)同时对服务器发起请求Webbench 每秒请求数测试-t 指定压测时间-c 指定并发用户数(非请求数)-f 零等待服务器响应如下图做了简单的性能压测.可以看出webbench模拟一个client相当于每秒有264.8个请求,，如果并发压-c 1000clients,相当于每秒有1000*264.8请求~想想也是相当凶残~~Ab每秒请求数测试-n 设置请求总数-c 设置并发client数量如图可以看出ab模拟一个client相当于每秒有268.04有个请求,，如果并发压-c 1000clients,相当于每秒有1000*268.04请求ab/webbench如何模拟用户行为业界一直认为loadrunner是最专业的,ab,webbench相比之下哪里不专业了呢,测试同学的说法是webbench,ab只进行了2次握手就离开继续下一次新请求,这里个人觉得不靠谱,我们模拟一次测试验证下.这里有遇到一个问题,如果去抓取一个用户数据包.ab的功能就有大显身手了,只压一个数据包.ab -n 1 -c 1 -k /index.php同时在被压测机抓包http请求数据流解析抓包命令tcpdump -i eth0 ip host web-yv4 and port ! 22tcpdump -i eth0 ip host 9 and 12 and dst port 80tcpdump flagsTCP FlagFlag in tcpdumpFlag MeaningSYNsSyn packet, a session establishment request. The first part of any TCP connection.ACKackAck packet, used to acknowledge the receipt of data from the sender. May appear in conjunction with other flags.FINfFinish flag, used to indicate the sender’s intention to terminate the connection to the receiving host.RESETrIndicates the sender’s intention to immediately abort the existing connection.PUSHpSignals the immediate push of data from the sending host to the receiving host. For interactive applications such as telnet, the main issue is the quickest response time, which this “push” flag signals.URGENTurgUrgent data should take precedence over other data. For example, a Ctrl-C to terminate a FTP download.Placeholder.If the connection does not have a syn, finish, reset, or push flag set, this placefolder flag will be found