Ansible playbook的block

环境

控制节点：Ubuntu 22.04
Ansible 2.10.8
管理节点：CentOS 8

block

顾名思义，通过block可以把task按逻辑划分到不同的“块”里面，实现“块操作”。此外，block还提供了错误处理功能。

task分组

下面的例子，把3个task放到一个block里面。

创建文件 testBlock1.yml 如下：

---
- name: testBlock1hosts: alltasks:- name: My task 1block:- name: Part1debug:msg: "Hello Zhang San"- name: Part2debug:msg: "Hello Li Si"- name: Part3debug:msg: "Hello Wang Wu"when: 2 > 1

运行结果如下：

➜  temp ansible-playbook testBlock1.ymlPLAY [testBlock1] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Zhang San"
}TASK [Part2] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Li Si"
}TASK [Part3] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Wang Wu"
}PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=4    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

如果把block的判断条件 when: 2 > 1 改为 when: 2 == 1 ，则运行结果如下：

➜  temp ansible-playbook testBlock1.ymlPLAY [testBlock1] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
skipping: [192.168.1.55]TASK [Part2] ***************************************************************************************
skipping: [192.168.1.55]TASK [Part3] ***************************************************************************************
skipping: [192.168.1.55]PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=1    changed=0    unreachable=0    failed=0    skipped=3    rescued=0    ignored=0

可见，由于条件不满足，block里的3个task都没有执行。

错误处理

如果block里的某个task出错了，则后面的task不再运行。

创建文件 testBlock2.yml 如下：

---
- name: testBlock2hosts: alltasks:- name: My task 1block:- name: Part1debug:msg: "Hello Zhang San"- name: Part2command: /bin/false # will trigger an error- name: Part3debug:msg: "Hello Li Si"

运行结果如下：

➜  temp ansible-playbook testBlock2.ymlPLAY [testBlock2] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Zhang San"
}TASK [Part2] ***************************************************************************************
fatal: [192.168.1.55]: FAILED! => {"changed": true, "cmd": ["/bin/false"], "delta": "0:00:00.001904", "end": "2023-10-26 08:50:36.850526", "msg": "non-zero return code", "rc": 1, "start": "2023-10-26 08:50:36.848622", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=2    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0

可见，由于Part2出错，Part3并没有运行。

Ansible的错误处理有两个关键字：

rescue ：类似于 catch
always ：类似于 finally

先加上 always 看看效果：

---
- name: testBlock2hosts: alltasks:- name: My task 1block:- name: Part1debug:msg: "Hello Zhang San"- name: Part2command: /bin/false # will trigger an error- name: Part3debug:msg: "Hello Li Si"always:- name: Always do thisdebug:msg: "End End End"

➜  temp ansible-playbook testBlock2.ymlPLAY [testBlock2] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Zhang San"
}TASK [Part2] ***************************************************************************************
fatal: [192.168.1.55]: FAILED! => {"changed": true, "cmd": ["/bin/false"], "delta": "0:00:00.002734", "end": "2023-10-26 08:52:19.329781", "msg": "non-zero return code", "rc": 1, "start": "2023-10-26 08:52:19.327047", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}TASK [Always do this] ******************************************************************************
ok: [192.168.1.55] => {"msg": "End End End"
}PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=3    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0

可见，Part2出错， always 也会运行（当然，Part3不会运行）。

注意：加上 always ，failed仍然是1。

现在来试一下 rescue ：

---
- name: testBlock2hosts: alltasks:- name: My task 1block:- name: Part1debug:msg: "Hello Zhang San"- name: Part2command: /bin/false # will trigger an error- name: Part3debug:msg: "Hello Li Si"always:- name: Always do thisdebug:msg: "End End End"rescue:- name: Rescue tasksdebug:msg: "Something is wrong!"

运行结果如下：

➜  temp ansible-playbook testBlock2.ymlPLAY [testBlock2] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
ok: [192.168.1.55] => {"msg": "Hello Zhang San"
}TASK [Part2] ***************************************************************************************
fatal: [192.168.1.55]: FAILED! => {"changed": true, "cmd": ["/bin/false"], "delta": "0:00:00.001726", "end": "2023-10-26 09:00:01.785445", "msg": "non-zero return code", "rc": 1, "start": "2023-10-26 09:00:01.783719", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}TASK [Rescue tasks] ********************************************************************************
ok: [192.168.1.55] => {"msg": "Something is wrong!"
}TASK [Always do this] ******************************************************************************
ok: [192.168.1.55] => {"msg": "End End End"
}PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=4    changed=0    unreachable=0    failed=0    skipped=0    rescued=1    ignored=0

注意，always是在rescue之后运行的（Part3仍然不会运行）。

注：如果task没有出错， rescue 不会被触发。

rescue与handler

我们知道，当task运行成功，状态改变时，可以用 notify 来触发handler。但如果后续的task出错了，则当前task的handler并不会触发。如果有 rescue ，则handler仍然会被触发（在 always 之后）。在 rescue 中可以通过 meta: flush_handlers 来立即触发handler（在 always 之前）。

---
- name: testBlock2hosts: alltasks:- name: My task 1block:- name: Part1debug:msg: "Hello Zhang San"changed_when: truenotify: Run me even after an error- name: Part2command: /bin/false # will trigger an error- name: Part3debug:msg: "Hello Li Si"always:- name: Always do thisdebug:msg: "End End End"rescue:- name: Rescue tasks#debug:#  msg: "Something is wrong!"meta: flush_handlershandlers:- name: Run me even after an errordebug:msg: 'This handler runs even on error'

运行结果如下：

➜  temp ansible-playbook testBlock2.ymlPLAY [testBlock2] **********************************************************************************TASK [Gathering Facts] *****************************************************************************
ok: [192.168.1.55]TASK [Part1] ***************************************************************************************
changed: [192.168.1.55] => {"msg": "Hello Zhang San"
}TASK [Part2] ***************************************************************************************
fatal: [192.168.1.55]: FAILED! => {"changed": true, "cmd": ["/bin/false"], "delta": "0:00:00.002172", "end": "2023-10-26 09:11:26.530609", "msg": "non-zero return code", "rc": 1, "start": "2023-10-26 09:11:26.528437", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}RUNNING HANDLER [Run me even after an error] *******************************************************
ok: [192.168.1.55] => {"msg": "This handler runs even on error"
}TASK [Always do this] ******************************************************************************
ok: [192.168.1.55] => {"msg": "End End End"
}PLAY RECAP *****************************************************************************************
192.168.1.55               : ok=4    changed=1    unreachable=0    failed=0    skipped=0    rescued=1    ignored=0

本例中，Part1运行成功，且改变了状态，所以触发了handler Run me even after an error 。但由于Part2出错，如果没有 rescue ，则Part1的handler不会触发。加上 rescue 之后，就会触发Part1的handler。本例中加上了 meta: flush_handlers ，所以会立即触发handler。