Given that the machines are installed from scratch, logs should be small. Why we don't just copy all /var/log and journal contents ?
Done.
Another thing that we should also collect are the configuration files. This can help detect problems in the playbooks themselves when generating those files.
I've taken a basic approach to collect configuration from cluster nodes.
Please feel free to suggest changes to overall approach and nomenclature.
Done.
I've taken a basic approach to collect configuration from cluster nodes.
Please feel free to suggest changes to overall approach and nomenclature.