Ansible 安裝 Kafka Cluster

之前筆記 Ubuntu 安裝 Kafka KRaft cluster 紀錄到在 Ubuntu 上安裝 KRaft mode (不使用 ZooKeeper) 的 Kafka cluster，雖然內容大致算清楚，但畢竟在切換不同 host 時需要自行調整，彈性不足，加上團隊也早就使用 Ansible 來管理安裝腳本，所以連帶紀錄一下 Ansible 安裝的腳本

基本環境說明

1. Azure VM B2s (2 vcpu,4GiB memory) x3

   • 10.1.0.4
   • 10.1.0.5
   • 10.1.0.6

2. Linux (ubuntu 22.04)

3. Kafka 3.4.0

4. OpenJDK 11 JDK

5. ansible 2.10.2

Ansible 腳本

1. 專案結構

   • roles
     • kafka
       • defaults
         • main.yml
       • tasks
         • install.yml
         • main.yml
         • restart.yml
         • uninstall.yml
       • templates
         • kafka.j2
         • kafka.service.j2
       • vars
         • main.yml
   • inventories
     • dev.ini
   • ansible.cfg
   • main.yml

2. 完整內容

   • roles

     • kafka

       • defaults

         • main.yml

           ---
           purpose: upgrade
           kafka_replicationfactor: 3
           

       • tasks

         • install.yml

           ---
           - name: Install JDK
             become: true
             ansible.builtin.apt:
               name: openjdk-11-jdk
               state: present
                             
           - name: Add user kafka
             become: true
             shell: |
               groupadd --system kafka
               useradd -s /sbin/nologin --system -g kafka kafka
             ignore_errors: yes
                               
           - name: Create kafka data dir
             become: true
             file:
               path: "{{ item }}"
               state: directory
               mode: '1775'
               owner: kafka
               group: kafka
               recurse: true
             with_items:
               - "{{kafka_home}}_{{kafka_port}}"
               - "{{kafka_home}}_{{kafka_port}}/log"
                               
           - name: Downloading kafka
             become: true
             get_url:
               url: "{{ kafka_source_url }}"
               dest: "{{kafka_home}}_{{kafka_port}}/kafka.tgz"
               mode: '0755'
                               
           - name: Unzip kafka
             become: true
             shell: tar -xvzf kafka.tgz --strip 1
             args:
               chdir: "{{kafka_home}}_{{kafka_port}}"
                               
           - name: Config kafka for cluster & Setup server id
             become: true
             template: 
               src: templates/kafka.j2
               dest: "{{kafka_home}}_{{kafka_port}}/config/kraft/server.properties"
               owner: kafka
               group: kafka
                               
           - name: create cluster id
             become: true
             run_once: true
             shell: "{{kafka_home}}_{{kafka_port}}/bin/kafka-storage.sh random-uuid"
             register: cluster_id
                               
           - name: create get cluster id
             run_once: true
             set_fact:
               kafka_cluster_id: "{{ cluster_id.stdout }}"
                               
           - name: Create storage
             become: true
             shell: "{{kafka_home}}_{{kafka_port}}/bin/kafka-storage.sh format -t {{kafka_cluster_id}} -c {{kafka_home}}_{{kafka_port}}/config/kraft/server.properties"
                                   
           - name: Copy kafka's daemon config
             become: true
             template:
               src:  templates/kafka.service.j2
               dest: /etc/systemd/system/kafka_{{kafka_port}}.service
               mode: '0755'
               owner: kafka
               group: kafka
               backup: yes
                               
           - name: Start kafka service
             become: true
             systemd:
               name: kafka_{{kafka_port}}
               state: started
               enabled: yes
               daemon_reload: yes
           

         • main.yml

           ---
           - name: Uninstall kafka
             include_tasks: uninstall.yml
             when: 
               - purpose=='uninstall'
                               
           - name: Install kafka
             include_tasks: install.yml
             when: 
               - purpose=='install'
               - inventory_hostname in groups['kafka_all']
                               
           - name: Restart kafka
             include_tasks: restart.yml
             when: 
               - purpose=='restart'
               - inventory_hostname in groups['kafka_all']
           

         • restart.yml

           ---
           - name: restart kafka
             become: true
             systemd:
               name: kafka
               state: restarted
               enabled: yes
           

         • uninstall.yml

           ---
           - name: Stop kafka service
             become: true
             systemd:
               name: "{{item}}"
               state: stopped
               enabled: no
             ignore_errors: yes
             with_items: 
               - kafka_{{kafka_port}}
                               
           - name: Release kafka data dir
             become: true
             file:
               path: "{{item}}"
               state: absent
             with_items: 
               - "{{kafka_home}}_{{kafka_port}}"
               - "/etc/systemd/system/kafka_{{kafka_port}}.service"
                               
           - name: Systemd reread configs
             become: true
             systemd:
               daemon_reload: yes
           

         • templates

         • kafka.j2

           process.roles=broker,controller
           broker.id={{ host_index | int +1}}
           port={{kafka_port}}
                               
           controller.quorum.voters={{voters}}
           listeners=PLAINTEXT://:{{kafka_port}},CONTROLLER://:{{kafka_port|int+1}}
           advertised.listeners=PLAINTEXT://{{ip}}:{{kafka_port}}
           controller.listener.names=CONTROLLER
           listener.security.protocol.map=CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT
                               
           log.dirs={{kafka_home}}_{{kafka_port}}/log
           

         • kafka.service.j2

           [Unit]
           Description=Kafka
           After=network.target
             
           [Service]
           Type=simple
           User=kafka
           ExecStart=/bin/sh -c '{{kafka_home}}_{{kafka_port}}/bin/kafka-server-start.sh {{kafka_home}}_{{kafka_port}}/config/kraft/server.properties > {{kafka_home}}_{{kafka_port}}/log/kafka.log 2>&1'
           ExecStop={{kafka_home}}_{{kafka_port}}/bin/kafka-server-stop.sh
           Restart=on-abnormal
           
           [Install]
           WantedBy=multi-user.target
           

         • vars

         • main.yml

           ---
           host_index: "{{groups['kafka_all'].index(inventory_hostname)}}"
           voters: "{% set VOTER_ARR=[] %}{% for host in groups['kafka_all'] %}{% if   VOTER_ARR.insert(loop.index,(loop.index|string)+'@'+hostvars[host].ip+':'  +(hostvars[host].kafka_port|int+1)|string) %}{% endif %}{% endfor %}  {{VOTER_ARR|join(',')}}"
           kafka_version: 3.4.0
           kafka_partitions: 3
           kafka_source_url: "https://downloads.apache.org/kafka/{{kafka_version}}/  kafka_2.13-{{kafka_version}}.tgz"
           

   • inventories

     • dev.ini

       ansible_host 是外網 ip 而 ip 則為內網 ip，如果 ansible client 是在內網，可以簡為 ansible_host 就好

       [all]
       kafka1 ansible_ssh_host=10.1.0.4 ansible_ssh_user=yowko ansible_ssh_pass=pass123 ip=10.1.0.4 ansible_become_password=pass.123
       kafka2 ansible_ssh_host=10.1.0.5 ansible_ssh_user=yowko ansible_ssh_pass=pass123 ip=10.1.0.5 ansible_become_password=pass.123
       kafka3 ansible_ssh_host=10.1.0.6 ansible_ssh_user=yowko ansible_ssh_pass=pass123 ip=10.1.0.6 ansible_become_password=pass.123
                   
       [kafka_all]
       kafka1 kafka_port=9092 
       kafka2 kafka_port=9092 
       kafka3 kafka_port=9092 
                   
       [kafka_all:vars]
       kafka_home="/home/kafka"
       

   • ansible.cfg

     [defaults]
     host_key_checking = False
     roles_path = ./roles
     

   • main.yml

     - hosts: kafka_all
     any_errors_fatal: '{{ any_errors_fatal | default(true) }}'
     vars:
       purpose: 'install'
     roles:
       - kafka-cluster
     

3. 安裝

   ansible-playbook -i inventories/dev.ini main.yml
   

心得

大致上流程與之前筆記 Ubuntu 安裝 Kafka KRaft cluster 相同，主要是將原本的 script 改為 ansible 語法，另外原本 script 在處理多個 instance 較不彈性的部份也因為 ansible 特性的關係獲得有效解決

完整程式可以參考 yowko/ansible-kafka

參考資訊

1. Ansible 安裝 Redis Cluster
2. yowko/ansible-kafka