Recently I had an emergency call from a client who said “we restarted Postgres, and the database is not starting.” It turns out that the recently promoted slave, using WAL logging and pg_standby, was never removed from the slave group. … Continue reading →