利用ansible检测网络连通性(多个网段多IP)

        在云平台的运维中,有时会遇到某个节点的一个IP无法连通导致云平台故障,比较常见的是ceph的某个osd节点storage网络不通导致osd全部down.为了快速检测云平台全部网络的连通性,利用ansible自带的的fact,写了个playbook,特记录下.

1 每台主机三个网卡



2  脚本内容

---

- hosts: all

  #vars_prompt:  

  #  - name: share_user

  #    prompt: "input share_user"

  tasks:

    - block

        - name: restart acpid  service

          service: name=acpid state=restarted

        - name: get  the network connection  ip

          shell: |

              ping -c 2  "{{ hostvars[item[0]]['ansible_' + item[1]].ipv4.address }}"

          register: netinfo

          ignore_errors: yes

          with_nested:

            - "{{ groups['all'] }}"

            - ["eth0","eth1","eth2"]

        

        #- debug:

        #    var: netinfo

        - name: echo the no ping ip

          shell: echo "ip {{item.cmd}} is no ok" >>/root/noping.txt

          with_items:

            - "{{ netinfo.results }}"

          when: item.rc != 0

      delegate_to: localhost



3 测试 

      3.1 关掉minion1的eth2网卡



    

   3.2 执行检测脚本



  

    3.3 测试结果



补充一个playbook,利用fact的ansible_all_ipv4_addresses变量

---

- hosts: all

  become: yes

  become_user: root

  become_method: sudo

  tasks:

    - block:

        - name: check the net connection(ping)

          shell: ping -c 2 {{ item }}

          register: netResult

          ignore_errors: yes

          with_items:

            - "{{ ansible_all_ipv4_addresses }}"

          when: item !='240.0.0.1'          #增加过滤不想检测的IP

        - debug:

            var=netResult

        - name: get the no ping  ipadress

          shell: |

            echo "ip {{ item['item'] }} is unreachable" >>/root/noPing.txt

          with_items:

            "{{ netResult.results }}"

          when: item.item !='240.0.0.1' and item.rc !=0

      delegate_to: localhost