hastexo / ceph-checkup Goto Github PK

Hello,

Thanks for this tool.
I see multiple places where shell scripts are used instead of modules, is that intentional? Do you expect contributions to fix this?

next to that, I see some bad Ansible practices (local actions, not using YAML notation for modules, the use of ignore_errors without blocks, etc.). Is there any expectation to fix that, or do you need help fixing this?

Was this a PoC?

Whenever we're dealing with large log files, our use of the fetch module may trigger an out-of-memory condition as seen in ansible/ansible#11702 (possible reason explained here).

Might consider a direct invocation of scp (from shell or command) or using synchronize instead.

checkup_timestamp: "{{ lookup('pipe', 'date +%Y%m%d%H%M%SZ') }}" can be replaced with a fact (see ansible_date_time). The fact is gathered by default on each playbook run, hence this line is redundant.

If we integrate this with other repositories, running to all nodes doesn't make sense.

Would it make more sense to list all the groups for the playbook in https://github.com/citynetwork/ceph-checkup/blob/master/checkup.yml#L2 ?