乱读天书，不求甚解

周祎骏的个人云笔记

乱读天书，不求甚解

bash 3.4 编写并行运行的脚本

2016-05-05 15:11:03 88 0 0

admin

限制同时在后台跑n个job并观察他们的状态，当有进程消失就触发新的job（保持后台总共有n个job)，当后台job跑了超过一定时间后也kill 了它们。
```bash
[root@test ~]# cat test.sh
#!/bin/bash
max_back_process_number=$1
back_process_timeout=$2
back_cmd=$3
need_run_cmd_number=$4

wait_back_process()
{
    local max_back_process_number=$1
    local back_process_timeout=$2
    local back_process_number=`jobs -rp | wc -l `
    echo "back jobs $back_process_number"
    while [[ $back_process_number -ge $max_back_process_number ]];do
        for the_pid in `jobs -rp`;do
            local process_running_time=`ps -p $the_pid -o etime= | sed -r 's/(([0-9]*)-)?(([0-9]*):)?([0-9]*):([0-9]*)/0\2 * 86400 + 0\4 * 3600 + 0\5 * 60 + \6/g' | bc ` #获取进程运行的时间
            if [ -n "$back_process_timeout" ] && [ $process_running_time -ge $back_process_timeout ];then
                local kill_pids=`pstree -p $the_pid | egrep -o '([0-9]+)'` #获取这个进程以及它所有子进程的pid
                kill -9 $kill_pids 2> /dev/null
            fi
        done
        sleep 1
        back_process_number=`jobs -rp | wc -l `
    done
}

for i in `seq $need_run_cmd_number`
 do
    echo "fock a new process"
    eval "$back_cmd" &
    wait_back_process $max_back_process_number $back_process_timeout
done

wait_back_process 1 $back_process_timeout     #等待所有进程跑完
[root@test ~]# ./test.sh 10 10  "sleep 100" 1000
```

上一篇: bash 3.3 timeout 控制

下一篇: docker_ctr_crictl

文档导航