pdns-ansible
pdns-ansible copied to clipboard
Feat(postgresql): Add support for postgresql backend (including tests)
Hi,
I have a use case where I would like to use postgressql as a backend for pdns.
This PR finishes the work started by #104
I've made the following changes:
-
Minor changes to #104
-
Add tests to the postgresSQL ansible code (I've added tests similar to what is tested with MySQL backend)
- Added a PostgreSQL service to test creation of postgresql user/database/schemas
-
Added "throttle:1" to both MySQL and PostgreSQL user creation module. Occasionally the tasks failed and I'm assuming it's because of race-conditions. (For both MySQL and PostgreSQL, a user with the same name is created multiple times (foreach distro in the tests) and assigned permissions/privileges)
-
Added support for PostgreSQL's Host Based Authentication. This can be used when PowerDNS Authoritative server is installed on the same host as the PostgreSQL server. In that case, Ansible does not need a privileged PostgreSQL user to create a user for PowerDNS. It will simply use the unix socket and use
become "postgresql". This is unfortunately not covered by any tests.
You could have kept my commit. To me this is borderline copyright infringement.
@commonism Your contribution is licensed under the MIT license. It explicitly allows merging code without attribution.
You can discuss open source etiquette if you want, but legalities have nothing to do with it.
Hi @commonism,
Done
@commonism Your contribution is licensed under the MIT license. It explicitly allows merging code without attribution.
I thought this was MIT-0, but well, can you point me to somewhere to read up on this?
You can discuss open source etiquette if you want, but legalities have nothing to do with it.
He's settled things already.
I thought this was MIT-0, but well, can you point me to somewhere to read up on this?
Here is the LICENSE file in your contribution repository.
MIT-0 is "zero attribution", the license here is MIT. I was not aware MIT can be read as merge-without-attribution as well and was asking for an explanation of this.
@commonism I've created a discussion in a throwaway repo to continue the license talk - there's no need to spam this PR with irrelevant discussion (open source etiquette and all :wink:)