工作中经常会遇到端口占用导致服务无法启动的情况， 这时候通常需要三步来解决问题：

1. 找到占用端口的 程序pid(netstat -tunlp | grep pid)

2. 通过pid 查看并确定程序(ps aux | grep pid)

3.kill掉该程序，重新启动服务即可 (kill -9 pid)

示例如下：

如查看5140端口的情况，netstat -tunlp |grep 5140

# netstat -tunlp

Active Internet connections (only servers)

Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name

tcp 0 0 0.0.0.0:111 0.0.0.0:* LISTEN 4814/rpcbind

tcp 0 0 0.0.0.0:5908 0.0.0.0:* LISTEN 25492/qemu-kvm

tcp 0 0 0.0.0.0:6996 0.0.0.0:* LISTEN 22065/lwfs

tcp 0 0 192.168.122.1:53 0.0.0.0:* LISTEN 38296/dnsmasq

tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 5278/sshd

tcp 0 0 127.0.0.1:631 0.0.0.0:* LISTEN 5013/cupsd

tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN 5962/master

tcp 0 0 0.0.0.0:8666 0.0.0.0:* LISTEN 44868/lwfs

tcp 0 0 0.0.0.0:8000 0.0.0.0:* LISTEN 22065/lwfs

[root@gy01 jar]# netstat -unlp | grep 5140

udp 0 0 0.0.0.0:5140 0.0.0.0:* 26805/java

关于几个参数的含义

-t (tcp) 仅显示tcp相关选项

-u (udp)仅显示udp相关选项

-n 拒绝显示别名，能显示数字的全部转化为数字

-l 仅列出在Listen(监听)的服务状态

-p 显示建立相关链接的程序名

查看  进程

[root@gy01 jar]# ps -aux | grep 26805

root 20678 0.0 0.0 112676 976 pts/0 S+ 12:00 0:00 grep --color=auto 26805

flume 26805 4.5 9.8 29811516 13009992 ? Sl 12月03 117:25 /home/flume/software/java/bin/java -XX:-UseLoopPredicate -XX:CompileCommand=exclude,org.apache.flume.source.kafka.KafkaSourceManager$KafkaConsumerRunnable::run -Xms20g -Xmx20g -XX:MaxDirectMemorySize=512M -XX:+UseG1GC -Dcom.sun.management.jmxremote -Daudit.properties.conf=/home/flume/software/flume-1.6.1.6/bin/../conf/log_audit.properties -Dfield.map.conf=/home/flume/software/flume-1.6.1.6/bin/../conf/field_map.properties -Dflume.root.logger=INFO,LOGFILE -Dflume.kerberos.enabled=true -cp /home/flume/software/flume-1.6.1.6/conf:/home/flume/software/flume-1.6.1.6/lib/*:/home/flume/software/flume-1.6.1.6/plugins.d/elasticSearch-sink/lib/*:/home/flume/software/flume-1.6.1.6/plugins.d/kafka-source/lib/*:/home/flume/software/hadoop-2.7.2.4/etc/hadoop:/home/flume/software/hadoop-2.7.2.4/share/hadoop/common/lib/activation-1.1.jar:/home/flume/software/hadoop-2.7.2.4/share/hadoop/common/lib/apacheds-i18n-2.0.0-M15.jar:/home/flume/software/hadoop-2.7.2.4/share/hadoop/common/lib/apacheds-kerberos-codec-2.0.0-M15.jar:/home/flume/software/hadoop-2.7.2.4/share/hadoop/common/lib/api-asn1-api-1.0.0-M20.jar:/home/flume/software/hadoop-2.。。。。。。。。。。。。。。。。。。。。。。。。。。

最后kill该进程即可

[root@gy01 ~]# kill -9 20893

贴一个监测端口占用的脚本(来自：http://www.cnblogs.com/wangtao1993/p/6144183.html )

#!/usr/bin/env python

# -*- coding:utf-8 -*-

import socket, time, thread

socket.setdefaulttimeout(3) #设置默认超时时间

def socket_port(ip, port):

"""

输入IP和端口号，扫描判断端口是否占用

"""

try:

if port >=65535:

print u'端口扫描结束'

s=socket.socket(socket.AF_INET, socket.SOCK_STREAM)

result=s.connect_ex((ip, port))

if result==0:

lock.acquire()

print ip,u':',port,u'端口已占用'

lock.release()

except:

print u'端口扫描异常'

def ip_scan(ip):

"""

输入IP，扫描IP的0-65534端口情况

"""

try:

print u'开始扫描 %s' % ip

start_time=time.time()

for i in range(0,65534):

thread.start_new_thread(socket_port,(ip, int(i)))

print u'扫描端口完成，总共用时：%.2f' %(time.time()-start_time)

# raw_input("Press Enter to Exit")

except:

print u'扫描ip出错'

if __name__=='__main__':

url=raw_input('Input the ip you want to scan: ')

lock=thread.allocate_lock()

ip_scan(url)

执行结果如下

# python scan_port.py

Input the ip you want to scan: 20.0.208.112

开始扫描 20.0.208.112

20.0.208.112 : 111 端口已占用

20.0.208.112 : 22 端口已占用

20.0.208.112 : 8000 端口已占用

20.0.208.112 : 15996 端口已占用

20.0.208.112 : 41734 端口已占用

扫描端口完成，总共用时：9.38