Replication configuration

[elho]

This adds key configuration settings needed to set up replication.

Compared to the existing two pull requests adding server_id (and log_bin in one case), this one

• allows configuration of server_id, log_bin, binlog_format, max_binlog_size and expire_logs_days
• provides a default value for log_bin (at least for Debian based distros)
• also adds comments for the new parameters
• uses instance variable references ('@') in the templates

[apenney]

I've taken a look at this and for the most part everything is good - it had no specs at all however. I've made a first pass at this for you, you can check https://gist.github.com/apenney/5785765 for the patch. If you can integrate that, make any improvements you want, and update the PR I'd feel happier about including it.

My only cause for concern at this point is having $log_bin only set for Debian. This means we're going to end up problems on other distros and I'd really prefer it if you could fill in sensible defaults to the best of your knowledge. I can test the results on all the other distros to make sure it works.

[abraham1901]

#167

[elho]

I did not provide the $log_bin defaults for other distros, as I do not know them and do think, that wrong (as in different from what a user of the distro expects) defaults are a bad idea. I do agree, that the values for all distros supported by the module should be filled in and I'll happily do it myself, if someone tells me the default values the missing distros use.

[apenney]

I think in terms of defaults for centos we can just match debian for now as there doesn't seem to be any defaults. I'd rather we put in what we want for now and adjust it if users feel its wrong. That way we can just get this merged in before the PR drifts too far.

If you're willing this needs a quick rebase against master and the defaults for the other operating systems added (and tests updating). Assuming you can take that on I'll make sure this gets merged as soon as it's rebased.

[apenney]

Alternatively, having said all that, if you take a peek at #167 it seems to include pretty much all the replication stuff you have here plus more. If you're comfortable with that PR we could switch focus to merging that one and wrapping this one up.

[apenney]

I merged in #167, hopefully that covers this feature!