Audit usage of systemd unit masking
When some services are moved to new hosts due to rescheduling, we can't start them:
fatal: [host3]: FAILED! => {"changed": false, "msg": "Unable to enable service prometheus.service: Synchronizing state of prometheus.service with SysV service script with /lib/systemd/systemd-sysv-install.\nExecuting: /lib/systemd/systemd-sysv-install enable prometheus\nFailed to enable unit: Unit file /etc/systemd/system/prometheus.service is masked.\n"}
a theory is that this happens because we've disabled the service with masked: yes in a systemd Ansible task, and we do not have masked: no when re-enabling it. If this is the case, we should make sure to add masked: no whenever we enable a systemd unit.
Edited by ale