pcerny

I recently also hit this issue. My use case requires two timeouts, one for establishing connection with Vertica and one for waiting to query response. When the Vertica node is...

@sitingren It depends when you plan to drop support for Python 2. If you will follow Python 2 EOL date then it is not worth to fix it. > I...

We are fighting with this issue of having two masters in bitnami/postgresql-ha deployment too. The following two scenarios lead to the state in our case: - temporary network hiccup which...

Let me elaborate on statement "works well with 3 node cluster only" in relation to `repmgrFenceOldPrimary: true` setting. First our motivation for covering network hiccups. We have deployments also in...

We did not want to go with `failover_on_backend_error = on` in pgpool. It marks node as down in pgpool immediately when ping to PG node fails. Health check functionality is...

@javsalgar @rafariossaa do you plan to evaluate above mentioned solutions, pick some and put it into bitnami/postgresql-ha? We would love to have the issue addressed in upstream. I can help...

@rafariossaa I have created PRs to bitnami pgpool and postgresql-ha repositories to make my suggestion presented above easier to evaluate. With these two PRs applied, the only change to PG-HA...