Ensure there is always at least one PHP-FPM pool defined

[fullyint]

History

Occasionally, the Start php7.0-fpm service task will fail with the error below.

Job for php7.0-fpm.service failed because the control process exited with
error code. See "systemctl status php7.0-fpm.service" and "journalctl -xe"
for details.

#642 took a first step to prevent this error and the scenario as presented in discourse.roots.io/t/7444. Some of the reports in discourse.roots.io/t/7418 were probably related.

News

The error can still occur when the playbook is rerun after previously failing between these two tasks:

• php : Disable default pool
• wordpress-setup : Create WordPress php-fpm configuration file

Between these two tasks, no PHP-FPM pool defined. The systemctl status php7.0-fpm.service reveals ERROR: No pool defined. at least one pool section must be specified in config file.

Unfortunately the playbook occasionally fails between these two tasks, e.g., due to connectivity issues while downloading a dependency.

Replicate

To replicate the problem, add this task to the end of roles/php/tasks/main.yml:

- fail:
    msg: provision again and watch the 'Start php7.0-fpm service' task fail

Then run the server.yml twice and see the error in question on the second run.

Resolution

This PR resolves the issue by removing the default pool only after the new WP pool is created.

[swalkinshaw]

Makes sense