centos7 网卡时常down，求分析

bamqjurgg

测试服务器基础信息：

操作系统：CentOS7.0
服务器型号：PowerEdge T620
网卡型号：
06:00.0 Ethernet controller: Intel Corporation I350 Gigabit Network Connection (rev 01)
06:00.1 Ethernet controller: Intel Corporation I350 Gigabit Network Connection (rev 01)

网卡驱动版本：
driver: igb
version: 5.3.0-k
firmware-version: 1.56, 0x800009a1, 14.0.12
expansion-rom-version:
bus-info: 0000:06:00.0
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes
supports-priv-flags: no


网卡经常莫名奇妙的down掉，宿主机上装了kvm，尤其是在还原快照时网卡down或者直接系统自动重启了

其message相关信息为：

1. Jan 24 10:05:20 develop systemd: Started Network Manager Script Dispatcher Service.
   
2. Jan 24 10:05:20 develop dbus-daemon: dbus[711]: [system] Successfully activated service 'org.freedesktop.machine1'
   
3. Jan 24 10:05:20 develop dbus[711]: [system] Successfully activated service 'org.freedesktop.machine1'
   
4. Jan 24 10:05:20 develop systemd: Started Virtual Machine and Container Registration Service.
   
5. Jan 24 10:05:20 develop systemd-machined: New machine qemu-1-win2k8.
   
6. Jan 24 10:05:20 develop systemd: Unit iscsi.service cannot be reloaded because it is inactive.
   
7. Jan 24 10:05:20 develop systemd: Started Virtual Machine qemu-1-win2k8.
   
8. Jan 24 10:05:20 develop systemd: Starting Virtual Machine qemu-1-win2k8.
   
9. Jan 24 10:05:21 develop kvm: 1 guest now active
   
10. Jan 24 10:05:21 develop journal: libvirt version: 2.0.0, package: 10.el7_3.2 (CentOS BuildSystem <http://bugs.centos.org>, 2016-12-06-19:53:38, c1bm.rdu2.centos.org)
    
11. Jan 24 10:05:21 develop journal: hostname: develop.org
    
12. ----->注意的地方

复制代码

因为网卡时常down，故写了个简单的脚本当ping几次返回全是丢失时就重启网卡

1. #!/bin/bash
   
2. export PAHT=$PATH
   
3. set -x
   
4. test=`ping -c 5 192.168.8.254 | grep "100% packet loss" | wc -l`
   
5. test
   
6. echo $test
   
7. if [ $test = 0 ];
   
8. then
   
9.         echo `date` >> logs
   
10.         echo "The network is working" >> logs
    
11. else
    
12.         /usr/sbin/ifconfig em1 up
    
13. fi

复制代码

crond日志显示脚本正常执行

1. <blockquote>Jan 24 10:05:01 develop CROND[2479]: (root) CMD (/root/network.sh            > /dev/null 2>&1)

复制代码

但自写脚本的日志记录如下，从10点10分过后是一直执行 /usr/sbin/ifconfig em1 up 但是没有起作用

1. [root@develop ~]# cat logs  | grep -A 20 "Jan 24"
   
2. Tue Jan 24 10:05:05 CST 2017
   
3. The network is working
   
4. Tue Jan 24 10:10:05 CST 2017
   
5. The network is working
   
6. Sun Feb 5 08:55:05 CST 2017
   
7. The network is working
   
8. Sun Feb 5 09:00:05 CST 2017
   
9. The network is working
   
10. Sun Feb 5 09:05:05 CST 2017
    
11. The network is working
    
12. 
    

复制代码

附上内核相关crash

1. [root@develop crash]# pwd
   
2. /var/crash
   
3. [root@develop crash]# ll
   
4. total 0
   
5. drwxr-xr-x 2 root root 42 Jan 11 08:34 127.0.0.1-2017-01-10-19:34:32
   
6. drwxr-xr-x 2 root root 42 Jan 13 11:03 127.0.0.1-2017-01-12-22:02:59
   
7. drwxr-xr-x 2 root root 42 Jan 18 19:31 127.0.0.1-2017-01-18-06:31:27
   
8. drwxr-xr-x 2 root root 42 Jan 18 21:22 127.0.0.1-2017-01-18-08:22:22
   
9. drwxr-xr-x 2 root root 42 Jan 18 21:54 127.0.0.1-2017-01-18-08:54:20
   
10. drwxr-xr-x 2 root root 42 Jan 19 10:02 127.0.0.1-2017-01-18-21:02:10
    
11. drwxr-xr-x 2 root root 42 Jan 19 10:50 127.0.0.1-2017-01-18-21:50:43
    

复制代码

由于ssh无法连接，故手工重启服务器，求帮助，遇到这种情况如何分析，从上边的错误看出，可能时磁盘问题，网卡驱动也是手工重新装的最新版，centos7自带的本版太频繁down，换了新版本稍微稳定些