參考: http://www.51testing.com/html/28/116228-238978.html http://blog.chinaunix.net/uid-27571599-id-3473078.html 在bash中,使用後臺任務來實現任務的多進程化。在不加控制的模式下,無論有多少任務,所有都後臺執行。也就是說,在這種狀況下,有多少任務就有多少「進程」在同時執行。 實例一:正常腳本(腳本功能:查看一個文件中的IP列表循環測試主機連通性) [root@ha-drbd-nfs-01-120 ~]# cat ip.txt 192.168.0.24 192.168.0.25 192.168.0.26 192.168.0.27 192.168.0.28 192.168.0.29 192.168.0.40 192.168.0.41 192.168.0.42 192.168.0.43 192.168.0.81 [root@heartbeat-02-122 ~]# cat ping_host.sh #!/bin/bash array=$(cat ip.txt) #echo $array for host in ${array[@]} do { ping $host -c 10 -w 1 -i 0.01 -q >/dev/null 2>&1 ret=$? if [ $ret -eq 0 ];then echo "$host isok" else echo "$host isbad" fi } done [root@heartbeat-02-122 ~]# time sh ping_host.sh 192.168.0.24 is bad 192.168.0.25 is bad …………………….. 117.79.235.238 is ok 210.14.130.130 is ok 210.14.156.210 is ok real 0m30.688s user 0m0.099s sys 0m0.501s 實例二:」多進程」實現(沒法精確控制併發數量) [root@heartbeat-02-122 ~]# cat ping_host.sh #!/bin/bash array=$(cat ip.txt) #echo $array for host in ${array[@]} do { ping $host -c 10 -w 1 -i0.01 -q >/dev/null 2>&1 ret=$? if [ $ret -eq 0];then echo"$host is ok" else echo"$host is bad" fi }& done [root@heartbeat-02-122 ~]# time sh ping_host.sh real 0m0.092s user 0m0.006s sys 0m0.014s 注:就在上面基礎上多加了一個後臺執行&符號,此時應該是全部循環任務併發執行 問題:進程數目不可控制的狀況 實例三:多線程實現 本實例說明了一種用wait、read命令模擬多線程的一種技巧,此技巧每每用於多主機檢查,好比ssh登陸、ping等等這種單進程比較慢而不耗費cpu的狀況,還說明了多線程的控制。 [root@heartbeat-02-122 ~]# cat ping_host.sh #!/bin/bash tmp_fifofile="/tmp/$.fifo" mkfifo $tmp_fifofile exec 6<>$tmp_fifofile rm $tmp_fifofile thread=5 for ((i=0;i<$thread;i++));do echo done >&6 array=$(cat ip.txt) #echo $array for host in ${array[@]} do read -u6 { ping$host -c 10 -w 1 -i 0.01 -q >/dev/null 2>&1 ret=$? if [$ret -eq 0 ];then echo "$host is ok" else echo "$host is bad" fi echo >&6 } & done wait exec 6>&- exit 0 [root@heartbeat-02-122 ~]# time sh ping_host.sh 192.168.0.27 is bad 192.168.0.26 is bad 192.168.0.25 is bad ……………………………. 210.14.156.210 is ok 210.14.130.130 is ok 117.79.235.238 is ok real 0m6.247s user 0m0.056s sys 0m0.349s