如何使用Ansible wait_for检查多行命令状态？

Question

Ansible v2.4.0.0

我正在安装Gitlab-CE，在Ansible任务中运行以下命令。如您所见，某些进程已关闭，但最终会启动。

# gitlab-ctl status
run: gitlab-workhorse: 0s, normally up
run: logrotate: 1s, normally up
down: nginx: 0s, normally up
down: postgresql: 1s, normally up
run: redis: 0s, normally up
run: sidekiq: 0s, normally up
run: unicorn: 0s, normally up

如何编写Ansible wait_for任务来检查所有服务何时都处于运行状态？IOW我只想看到下一个任务就继续下一个任务

# gitlab-ctl status
run: gitlab-workhorse: 0s, normally up
run: logrotate: 1s, normally up
run: nginx: 0s, normally up
run: postgresql: 1s, normally up
run: redis: 0s, normally up
run: sidekiq: 0s, normally up
run: unicorn: 0s, normally up

Answer 1

您可以使用retries/ until方法：

- command: gitlab-ctl status
  register: cmd_res
  retries: 5
  until: cmd_res.stdout_lines | reject('search','^run') | list | count == 0

我们从开头的输出中删除任何行，run并对它们进行计数。仅当没有剩余的行时才继续执行下一个任务。