多个apache root进程 [英] Multiple apache root processes
问题描述
我今天注意到，当从我们的 Web 服务器发出请求时，速度相当慢.我开始研究它，我发现了大量 root 拥有的 apache 进程.
I noticed today that when making requests from our web server, things were rather slow. I started looking into it and I've found a load of root owned apache processes.
我不确定这实际上是导致事情变慢的原因，但无论如何，它看起来并不好.
I don't know for sure that this is actually what's causing things to be slow, but none the less, it doesn't look good.
问题是，我不知道从这里做什么?我如何找出为什么有这么多根进程?有人可以推荐一组测试吗?我试过对其中的一些进行 stracing，它们似乎在做某事，但 strace 的输出超出了我的能力.
problem is, I don't know what to do from here? How do I find out why there are so many root processes? Could some recommend a set of tests? I've tried stracing a few of them, and they appear to be doing something, but the output of strace is beyond me.
root 30918 1.8 1.3 84284 52296 ? Ss 14:11 0:01 /usr/sbin/apache2 -k restart
root 30919 0.0 1.1 84420 45612 ? S 14:11 0:00 /usr/sbin/apache2 -k restart
root 30920 0.0 1.1 84420 45604 ? S 14:11 0:00 /usr/sbin/apache2 -k restart
root 30921 0.0 1.1 84420 45612 ? S 14:11 0:00 /usr/sbin/apache2 -k restart
root 30922 0.1 1.1 84420 45612 ? S 14:11 0:00 /usr/sbin/apache2 -k restart
root 30923 0.0 1.1 84420 45612 ? S 14:11 0:00 /usr/sbin/apache2 -k restart
www-data 30926 6.6 1.5 104964 61336 ? S 14:12 0:03 /usr/sbin/apache2 -k restart
root 30930 0.1 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30933 0.0 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30935 0.0 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30936 0.0 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30937 0.0 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30938 0.0 1.1 84420 45616 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30961 0.0 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30989 0.0 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 30990 0.0 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 31011 0.1 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 31013 0.1 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 31014 0.0 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31175 2.5 1.5 104168 60524 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31189 2.3 1.4 102360 58920 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31190 1.5 1.4 101904 58356 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31191 0.3 1.1 84556 46760 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31192 1.4 1.4 101916 58384 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
www-data 31193 1.5 1.4 101916 58376 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
root 31240 0.1 1.1 84420 45612 ? S 14:12 0:00 /usr/sbin/apache2 -k restart
这是其中一个进程的 strace 输出示例.
This is an example of the output from strace from one of the processes.
--- SIGCHLD (Child exited) @ 0 (0) ---
read(6, 0xff87f6ef, 1) = -1 EAGAIN (Resource temporarily unavailable)
getuid32() = 0
close(17) = 0
gettimeofday({1354109303, 670988}, NULL) = 0
semop(5668864, {{0, -1, SEM_UNDO}}, 1) = 0
accept(4, {sa_family=AF_INET, sin_port=htons(48107), sin_addr=inet_addr("192.168.16.12")}, [16]) = 17
fcntl64(17, F_GETFD) = 0
fcntl64(17, F_SETFD, FD_CLOEXEC) = 0
semop(5668864, {{0, 1, SEM_UNDO}}, 1) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf74a2768) = 1949
waitpid(1949, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 1949
--- SIGCHLD (Child exited) @ 0 (0) ---
read(6, 0xff87f6ef, 1) = -1 EAGAIN (Resource temporarily unavailable)
getuid32() = 0
close(17) = 0
gettimeofday({1354109305, 724358}, NULL) = 0
semop(5668864, {{0, -1, SEM_UNDO}}, 1) = 0
accept(4, {sa_family=AF_INET, sin_port=htons(48132), sin_addr=inet_addr("192.168.16.12")}, [16]) = 17
fcntl64(17, F_GETFD) = 0
fcntl64(17, F_SETFD, FD_CLOEXEC) = 0
semop(5668864, {{0, 1, SEM_UNDO}}, 1) = 0
clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xf74a2768) = 1974
waitpid(1974, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0) = 1974
--- SIGCHLD (Child exited) @ 0 (0) ---
我已经禁用了 mods-enabled 中的所有模块，除了 auth、env、siteenv 和 alias 等基本模块并启动了服务器.在这种情况下，我仍然得到 6 个 root apache 进程和 1 个 www-data 拥有的 apache 进程.
I've disabled all of the modules in mods-enabled except for essential ones like auth, env, siteenv and alias and started the server. In this case I still get 6 root apache processes and 1 www-data owned apache process.
我已确保所有模块都是 up2date.
I've made sure all the modules are up2date.
日志中没有明显错误.
配置跟随；
ServerRoot "/etc/apache2"
LockFile /var/lock/apache2/accept.lock
PidFile ${APACHE_PID_FILE}
Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
<IfModule mpm_worker_module>
StartServers 2
MaxClients 150
MinSpareThreads 25
MaxSpareThreads 75
ThreadsPerChild 25
MaxRequestsPerChild 0
</IfModule>
User ${APACHE_RUN_USER}
Group ${APACHE_RUN_GROUP}
AccessFileName .htaccess
<Files ~ "^\.ht">
Order allow,deny
Deny from all
</Files>
DefaultType text/plain
HostnameLookups Off
ErrorLog /var/log/apache2/error.log
LogLevel warn
Include /etc/apache2/mods-enabled/*.load
Include /etc/apache2/mods-enabled/*.conf
Include /etc/apache2/httpd.conf
Include /etc/apache2/ports.conf
LogFormat "%v:%p %h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" vhost_combined
LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"" combined
LogFormat "%h %l %u %t \"%r\" %>s %b" common
LogFormat "%{Referer}i -> %U" referer
LogFormat "%{User-agent}i" agent
CustomLog /var/log/apache2/other_vhosts_access.log vhost_combined
Include /etc/apache2/conf.d/
Include /etc/apache2/sites-enabled/
模块中编译的有:在模块中编译:核心文件mod_log_config.cmod_logio.citk.chttp_core.cmod_so.c
The compiled in modules are: Compiled in modules: core.c mod_log_config.c mod_logio.c itk.c http_core.c mod_so.c
所以我现在只运行 mpm_worker 配置.
So I'm only running the mpm_worker config now.
DEBUG UPDATER
When I restart apache, and ps, I get something like this;
root 26921 0.5 1.3 80008 52452 ? Ss 21:27 0:02 /usr/sbin/apache2 -k start
root 27114 0.0 1.1 80144 44804 ? S 21:34 0:00 /usr/sbin/apache2 -k start
root 27115 0.0 1.1 80144 44820 ? S 21:34 0:00 /usr/sbin/apache2 -k start
root 27116 0.0 1.1 80144 44804 ? S 21:34 0:00 /usr/sbin/apache2 -k start
root 27117 0.0 1.1 80144 44804 ? S 21:34 0:00 /usr/sbin/apache2 -k start
root 27119 0.0 1.1 80144 44804 ? S 21:34 0:00 /usr/sbin/apache2 -k start
如果我将 LogLevel 设置为调试并重新启动，那么我会看到来自 mod_proxy 的这些消息
If I put LogLevel to debug and restart, then I see these messages from mod_proxy
[Thu Nov 29 21:34:01 2012] [info] Server built: Sep 9 2012 21:17:36
[Thu Nov 29 21:34:01 2012] [debug] itk.c(1100): AcceptMutex: sysvsem (default: sysvsem)
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1818): proxy: grabbed scoreboard slot 0 in child 27115 for worker proxy:reverse
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1818): proxy: grabbed scoreboard slot 0 in child 27114 for worker proxy:reverse
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1934): proxy: initialized single connection worker 0 in child 27115 for (*)
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1837): proxy: worker proxy:reverse already initialized
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1934): proxy: initialized single connection worker 0 in child 27114 for (*)
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1818): proxy: grabbed scoreboard slot 0 in child 27117 for worker proxy:reverse
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1837): proxy: worker proxy:reverse already initialized
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1934): proxy: initialized single connection worker 0 in child 27117 for (*)
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1818): proxy: grabbed scoreboard slot 0 in child 27119 for worker proxy:reverse
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1837): proxy: worker proxy:reverse already initialized
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1934): proxy: initialized single connection worker 0 in child 27119 for (*)
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1818): proxy: grabbed scoreboard slot 0 in child 27116 for worker proxy:reverse
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1837): proxy: worker proxy:reverse already initialized
[Thu Nov 29 21:34:01 2012] [debug] proxy_util.c(1934): proxy: initialized single connection worker 0 in child 27116 for (*)
[Thu Nov 29 21:36:20 2012] [notice] SIGHUP received. Attempting to restart
注意 pid 匹配.但是，如果我禁用 mod_proxy，那么这些消息就会消失，但我仍然会启动相同数量的 root 进程，所以我相信这是一个症状而不是原因.
Notice the pids match. However, if I disable mod_proxy, then these message disappear, but I still get the same number of root processes starting, so I believe this is a symptom not a cause.
推荐答案
这对于 Apache 来说是绝对正常的.每个进程一次处理一个请求.因此，如果只有一个进程(称为 worker)，那么如果有很多用户，它会非常慢.
This is absolutely normal for Apache. Each process processes one request at a time. So if there was only one process (it is called worker) then it would be really slow if there are lots of users.
我看到的问题是这些不应该是 root 拥有的进程.根据您的平台，它应该有自己的用户.就像在 Debian 中，用户是 www-data.那么只有一个进程由 root 拥有，其余的将由该用户拥有.
The issue I see is that these should not be root owned processes. Depending on your platform it should have it's own user. Like in Debian user would be www-data. Then only one process would be owned by root and rest would be owned by that user.
然而，速度是由几个因素决定的 - 硬件、网络服务器和网络应用程序.
However speed is defined by several factors - hardware, web server, and web application.
确保您运行的硬件符合要求(足够的内存和 CPU)
Make sure that hardware you are running on fits requirements (enough ram and CPU)
如果硬件能力差，则减少工人数量，如果超级好，则增加工人数量.
Lower number of workers in case of poor hardware capabilities or increase if it is super good.
确保 Web 应用程序(如果有，通常是 php 应用程序)不是性能瓶颈.
Make sure that web application (if there's is one, and often it is php app) is not a bottleneck for performance.
PS:抱歉格式不正确，从电话里硬着头皮输入.
PS: sorry for poor formatting, typed clamsily from phone.
这篇关于多个apache root进程的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
