add keeper flag to generate pgHBA config on restarts

[FarhanSajid1]

Changes:

• added in additional flag to control whether we want to skip the rendering of the hba configuration file defaulting to true to preserve all existing behavior

Reasons: Currently in our setup, there is synchronous replication which means when the cluster is brought down stolon doesn't regenerate the custom pgHBA files from the postgresql.json file. This could potentially lead to missing auth entries in the pg_hba.conf file, and lead to a situation where the cluster never comes up healthy Example: no pg_hba.conf entry for host \"127.0.0.1\", user \"ns1\", database \"ns1\", SSL off

There should be a flag to trigger the generation of this file in this situation to on/off, but please let us know if that doesn't make sense or there's a flaw in that logic that would result in unwanted behavior.

[sgotti]

@FarhanSajid1 If you're encountering an issue please open a dedicated issue with a way to reproduce it (or better a test case) before opening a PR so we'll be able to understand and find a fix. I don't think that there's the need to add a flag to fix it.

[pgray]

@sgotti we've recently had some users fully stop (all at once) a cluster of 3 nodes. on startup (sometimes), the cluster comes up, goes healthy, but then the old master is re-elected and this setting prevents it from becoming the cluster master (or at least blocks our postgres connection into it). A subsequent kick fixes it and it re-renders the config.

We'll get the issue opened up.

[FarhanSajid1]

@FarhanSajid1 If you're encountering an issue please open a dedicated issue with a way to reproduce it (or better a test case) before opening a PR so we'll be able to understand and find a fix. I don't think that there's the need to add a flag to fix it.

@sgotti Related issue: https://github.com/sorintlab/stolon/issues/792

[sgotti]

Closing since this isn't really needed.