如何不通过GUI在OpenDaylight 16.02控制器中运行脚本配置流表？

针对你在ODL 16.02控制器端用脚本增改流表的需求，以下是两种直接可行的方案，无需依赖GUI：

方法1：RESTAPI + 脚本（Python/curl） #

ODL原生支持RESTCONF接口，可直接在控制器主机上编写脚本调用本地REST端点，实现流表的增删改查。

示例：Python脚本添加双向转发流表（实现主机互通） #

假设拓扑为h1(10.0.0.1) -> s1 -> h2(10.0.0.2)，以下脚本可添加双向转发流表：

import requests
from requests.auth import HTTPBasicAuth

# ODL基础配置
ODL_ADDR = "localhost:8181"
AUTH = HTTPBasicAuth("admin", "admin")
RESTCONF_FLOW_URL = f"http://{ODL_ADDR}/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow"

# h1到h2的流表条目
flow_h1_to_h2 = {
    "flow": [
        {
            "id": "1",
            "match": {"ipv4-source": "10.0.0.1/32", "ipv4-destination": "10.0.0.2/32"},
            "instructions": {
                "instruction": [
                    {
                        "order": "0",
                        "apply-actions": {
                            "action": [{"order": "0", "output-action": {"output-node-connector": "2"}}]
                        }
                    }
                ]
            },
            "priority": "100",
            "table_id": "0"
        }
    ]
}

# h2到h1的流表条目
flow_h2_to_h1 = {
    "flow": [
        {
            "id": "2",
            "match": {"ipv4-source": "10.0.0.2/32", "ipv4-destination": "10.0.0.1/32"},
            "instructions": {
                "instruction": [
                    {
                        "order": "0",
                        "apply-actions": {
                            "action": [{"order": "0", "output-action": {"output-node-connector": "1"}}]
                        }
                    }
                ]
            },
            "priority": "100",
            "table_id": "0"
        }
    ]
}

# 发送请求添加流表
def add_flow(flow_data):
    headers = {"Content-Type": "application/json"}
    resp = requests.post(RESTCONF_FLOW_URL, auth=AUTH, headers=headers, json=flow_data)
    print(f"添加流表状态码: {resp.status_code}, 响应: {resp.text}")

add_flow(flow_h1_to_h2)
add_flow(flow_h2_to_h1)

使用说明：将脚本放在ODL控制器主机上，安装requests库（pip install requests）后运行，需根据实际拓扑修改节点ID（openflow:1）、端口号（1/2）、IP地址等参数。

示例：Shell脚本用curl修改流表 #

如果偏好shell脚本，可直接用curl调用RESTCONF：

# 添加h1到h2的流表
curl -u admin:admin -X POST \
  http://localhost:8181/restconf/config/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow/1 \
  -H "Content-Type: application/json" \
  -d '{
    "flow": {
        "id": "1",
        "match": {"ipv4-source": "10.0.0.1/32", "ipv4-destination": "10.0.0.2/32"},
        "instructions": {
            "instruction": [
                {
                    "order": "0",
                    "apply-actions": {"action": [{"order": "0", "output-action": {"output-node-connector": "2"}}]}
                }
            ]
        },
        "priority": "100",
        "table_id": "0"
    }
}'

方法2：ODL Karaf命令行脚本 #

ODL的Karaf控制台支持批量执行命令，可将流表操作写成.karaf脚本，直接在控制器上运行。

示例Karaf脚本（flow_config.karaf） #

# 添加双向转发流表
openflow:add-flow openflow:1 0 1 100 ipv4,src=10.0.0.1/32,dst=10.0.0.2/32 output=2
openflow:add-flow openflow:1 0 2 100 ipv4,src=10.0.0.2/32,dst=10.0.0.1/32 output=1

# 验证流表是否添加成功
openflow:show-flows openflow:1

运行方式：进入ODL的bin目录，执行./karaf run flow_config.karaf，或在Karaf控制台内用source flow_config.karaf加载执行。

流量监控脚本思路 #

通过RESTAPI定期获取流表统计数据，即可实现流量监控：

• 调用http://localhost:8181/restconf/operational/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow接口，提取每个流表条目的packet-count（数据包数）和byte-count（字节数）字段。
• 编写Python脚本定时请求该接口，解析并输出监控数据，或存入日志。

简单监控脚本示例 #

import requests
import time
from requests.auth import HTTPBasicAuth

ODL_STATS_URL = "http://localhost:8181/restconf/operational/opendaylight-inventory:nodes/node/openflow:1/flow-node-inventory:table/0/flow"
AUTH = HTTPBasicAuth("admin", "admin")

def get_flow_stats():
    headers = {"Accept": "application/json"}
    resp = requests.get(ODL_STATS_URL, auth=AUTH, headers=headers)
    if resp.status_code == 200:
        flows = resp.json()["flow-node-inventory:flow"]
        for flow in flows:
            print(f"流表ID {flow['id']}: 数据包数={flow['packet-count']}, 字节数={flow['byte-count']}")
    else:
        print(f"获取统计失败，状态码: {resp.status_code}")

# 每5秒刷新一次监控数据
while True:
    get_flow_stats()
    time.sleep(5)

注意事项 #

• 确保ODL已安装openflowplugin和restconf特性，在Karaf控制台执行feature:list | grep openflow和feature:list | grep restconf确认，未安装则用feature:install odl-openflowplugin-all odl-restconf安装。
• 流表的优先级、匹配字段需根据实际拓扑调整，避免流表冲突。
• RESTCONF认证信息需与你的ODL配置一致。

内容的提问来源于stack exchange，提问作者Mohammad Hajyghasem