Infrastructure

I finish the Ansible rules deployment, which needs to be reviewed:
https://github.com/bmbouter/pulp-infra-ansible/pull/4

I discovered the Redmine plugin is now back to the original code (and not the Pulp fork). I then adapted the configuration accordingly.

I stopped the bot, synced the data, and started the new one. The bot connects without error.

I tested the Redmine plugin and it fails with a JSON error. Involved libraried being the same, I tested the original bot and it failed saying two plugins where claiming the command, so it did not work. I did some debug and found out the rest library did not follow the plan.io HTTP->HTTPS redirection and this led to an empty string and JSON decoding failure. it is not secure anyway so I changed the URL to HTTPS and it works fine now.

I also switched the RSS feed to HTTPS, as it is available.

I also switched on the certificate validation.

@dkliban could you check the bugtriage/meetbot part please?

Also for the record we're using packaged libs only, no pip install or env. python2-restkit was missing so I backported it along with a few deps in the OSAS RPM repository. We already are using this repository for other reasons, so no change here.

The meetbot plugin having a replay command, they suggest to configure using a specific configuration file directly into the plugin directory; I did not see that coming. As we're using the packaged version, this is not possible. Pulp does not use the replay command as logs are published via web, so I used the supybot-style config. I updated the Ansible PR accordingly.

Also, the PulpTriage plugin was still using the PulpRedmine fork internally, but as the original one is now used and the old one not needed anymore, I created a PR to fix this.