运维宝典

## 安装问题
### sed修改/etc/hosts报错
```shell
[root@localhost install]# sed -i "/ $name\$/d" /etc/hosts
sed: cannot rename /etc/sedPCWJ83: Device or resource busy
```
## 重启后neutron-server无法启动
#### 现象

```shell
Jun 21 15:48:15 node23140 neutron-server: config.init(sys.argv[1:])
Jun 21 15:48:15 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/neutron/common/config.py", line 78, in init
Jun 21 15:48:15 node23140 neutron-server: **kwargs)
Jun 21 15:48:15 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/oslo_config/cfg.py", line 2504, in __call__
Jun 21 15:48:15 node23140 neutron-server: raise ConfigFilesNotFoundError(self._namespace._files_not_found)
Jun 21 15:48:15 node23140 neutron-server: oslo_config.cfg.ConfigFilesNotFoundError: Failed to find some config files: /etc/neutron/plugin.ini
Jun 21 15:48:15 node23140 systemd: neutron-server.service: main process exited, code=exited, status=1/FAILURE
Jun 21 15:48:15 node23140 systemd: Failed to start OpenStack Neutron Server.
Jun 21 15:48:15 node23140 systemd: Unit neutron-server.service entered failed state.
Jun 21 15:48:15 node23140 systemd: neutron-server.service failed.
Jun 21 15:48:15 node23140 systemd: neutron-server.service holdoff time over, scheduling restart.
Jun 21 15:48:15 node23140 systemd: Stopped OpenStack Neutron Server.
Jun 21 15:48:15 node23140 systemd: Starting OpenStack Neutron Server...
Jun 21 15:48:16 node23140 neutron-server: Traceback (most recent call last):
Jun 21 15:48:16 node23140 neutron-server: File "/usr/bin/neutron-server", line 10, in <module>
Jun 21 15:48:16 node23140 neutron-server: sys.exit(main())
Jun 21 15:48:16 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/neutron/cmd/eventlet/server/__init__.py", line 19, in main
Jun 21 15:48:16 node23140 neutron-server: server.boot_server(wsgi_eventlet.eventlet_wsgi_server)
Jun 21 15:48:16 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/neutron/server/__init__.py", line 40, in boot_server
Jun 21 15:48:16 node23140 neutron-server: _init_configuration()
Jun 21 15:48:16 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/neutron/server/__init__.py", line 30, in _init_configuration
Jun 21 15:48:16 node23140 neutron-server: config.init(sys.argv[1:])
Jun 21 15:48:16 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/neutron/common/config.py", line 78, in init
Jun 21 15:48:16 node23140 neutron-server: **kwargs)
Jun 21 15:48:16 node23140 neutron-server: File "/usr/lib/python2.7/site-packages/oslo_config/cfg.py", line 2504, in __call__
Jun 21 15:48:16 node23140 neutron-server: raise ConfigFilesNotFoundError(self._namespace._files_not_found)
Jun 21 15:48:16 node23140 neutron-server: oslo_config.cfg.ConfigFilesNotFoundError: Failed to find some config files: /etc/neutron/plugin.ini
Jun 21 15:48:16 node23140 systemd: neutron-server.service: main process exited, code=exited, status=1/FAILURE
Jun 21 15:48:16 node23140 systemd: Failed to start OpenStack Neutron Server.
Jun 21 15:48:16 node23140 systemd: Unit neutron-server.service entered failed state.
```
#### 原因

`/etc/neutron/目录下缺少指向/etc/neutron/plugins/ml2/ml2_conf.ini文件的名字为plugin.ini软连接。`

#### 解决方法

1. 创建软连接

```shell
ln -s /etc/neutron/plugins/ml2/ml2_conf.ini /etc/neutron/plugin.ini
```
2. 启动neutron-server服务

```shell
systemctl start neutron-server
```
## 安装后，计算节点状态是“未知”
#### 原因分析：

- 首先确定ndoutils数据库中的nagios_hosts和nagios_hoststat表中没有数据。这表中的数据是alarm.py写入的
- alarm.py是guard.sh调用的
- alarm.py需要从monitor数据库中的usedata_host表读取节点信息
- usedata_host表是manager进程写入的
#### 规避方法：

1. 重启manager进程
2. 直接手动执行alarm.py
## 计算节点上的host进程启动报错
#### 现象

```shell
host begin to start
MulticastRecvInit:setsockopt error
the error is19
```

#### 原因

计算节点上没有到广播地址ip route get 224.0.2.1的路由：

```shell
[root@computer1 var]# ip route get 224.0.2.1
RTNETLINK answers: Network is unreachable
```
#### 解决方法

1. 增加默认路由：

```shell
ip route add default via 11.0.0.1 dev br0
```
2. 重启控制节点上的manager进程

3. 重启计算节点上的host进程

## 页面没有监控数据
#### 原因1
- 计算节点上的host进程未上报数据，原因是host进程未收到控制节点manager进程发送的udp广播消息

- 进一步发现，控制节点上的默认路由是从另外一个192.168.2.0/24网段的网口

```shell
  [root@node23146 xfye]# ip route
  default via 192.168.2.1 dev brq88dc3d44-0d metric 99
  default via 11.0.0.1 dev em1 proto static metric 101
  11.0.0.0/16 dev em1 proto kernel scope link src 11.0.23.146 metric 101
  192.168.2.0/24 dev brq88dc3d44-0d proto kernel scope link src 192.168.2.101
```
#### 解决方法

- 删除192.168.2.0/24默认路由
#### 原因2
1502或者1523的udp端口被占用。

```shell
[root@computer3 host]# netstat -anp | grep 1523
udp     2304      0 0.0.0.0:1523            0.0.0.0:*                           9439/bash[root@computer3 host]# netstat -anp | grep 1502
udp   426240      0 0.0.0.0:1502            0.0.0.0:*                           9439/bash[root@computer3 host]# kill -9 9439
```
#### 解决方法

杀死占用1502和1523端口的进程，以及host进程。然后再次启动host程序。

## 使用问题
### 资源组
#### 剩余资源如何计算
`创建资源组时的可用CPU = (所有计算节点cpu个数 * 超分比例) - (所有资源组的当前CPU上限)`