为什么启动后台程序需要使用nohup

在linux服务器上,启动程序时在尾部添加&,可以把程序丢到后台运行。

当终端退出后,该终端启动的后台程序也可能会停止运行。

原因是终端退出后,该终端启动的后台程序会收到SIGHUP信号。如果后台程序内部没有处理SIGHUP信号,SIGHUP信号的默认处理行为是停止程序运行。

为什么后台程序可能会停止运行我用了可能两个字呢?因为我在实验时发现这和终端环境有关系。

当我使用一台终端配置的比较完善的mac电脑(安装了iterm,zsh,tmux等一堆乱七八糟的东西)去操作远程服务器时,我发现退出终端后,终端启动的后台程序并没有退出。为了进一步验证,我在关闭终端后,重新打开终端,在本地使用netstat和ps命令查看,发现和远程服务器的连接依然保持存在。

而当我使用一台比较原始的mac电脑去操作远程服务器时,我发现退出终端后,终端启动的后台程序也停止了。

为了避免终端退出时后台程序也停止运行,可选择使用如下几种方式:

  • 在后台程序内部编写处理SIGHUP信号的代码
  • 将后台程序编写为daemon守护进程
  • 使用nohup命令启动我们的后台程序

nohup官网地址: http://www.gnu.org/software/coreutils/manual/html_node/nohup-invocation.html

官网的简单介绍:
nohup runs the given command with hangup signals ignored, so that the command can continue running in the background after you log out.

关于Linux进程组、会话、守护进程、终端的资料:

我的测试程序,以及一些命令:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
#include <signal.h>
#include <stdio.h>

void func(int sig) {
// 通过 man signal,可以看到SIGHUP对应的信号值为1
// 1 SIGHUP terminate process terminal line hangup
fprintf(stderr, "sig:%d\n", sig);
fflush(stderr);
exit(0);
}

int main() {
int i = 0;
// 程序内部处理SIGHUP
signal(SIGHUP, func);
for (;;) {
fprintf(stderr, "%d\n", i++);
fflush(stderr);
sleep(1);
}
return 0;
}
1
2
3
4
5
6
7
8
9
10
11
12
# 远程
$gcc main.c
$./a.out
$./a.out 2> stderr.txt &
$nohup ./a.out 2> stderr.txt &
$jobs
$ps -ef | grep a.out
$tail -f stderr.txt

# 本地
$ps -ef | grep ssh
$netstat -an | grep 22

本文完,作者yoko,尊重劳动人民成果,转载请注明原文出处: https://pengrl.com/p/20022/

0%