[MISC] Improve TLS stability

src/charm.py

@@ -1724,6 +1724,12 @@ def update_config(self, is_creating_backup: bool = False) -> bool:
             return True
 
         if not self._patroni.member_started:
+            if enable_tls:
+                logger.debug(
+                    "Early exit update_config: patroni not responding but TLS is enabled."
+                )
+                self._handle_postgresql_restart_need(True)
+                return True
             logger.debug("Early exit update_config: Patroni not started yet")
             return False
 
@@ -1780,8 +1786,14 @@ def _validate_config_options(self) -> None:
 
     def _handle_postgresql_restart_need(self, enable_tls: bool) -> None:
         """Handle PostgreSQL restart need based on the TLS configuration and configuration changes."""
-        restart_postgresql = self.is_tls_enabled != self.postgresql.is_tls_enabled()
-        self._patroni.reload_patroni_configuration()
+        if self._can_connect_to_postgresql:
+            restart_postgresql = self.is_tls_enabled != self.postgresql.is_tls_enabled()
+        else:
+            restart_postgresql = False
+        try:
+            self._patroni.reload_patroni_configuration()
+        except Exception as e:
+            logger.error(f"Reload patroni call failed! error: {e!s}")
         # Wait for some more time than the Patroni's loop_wait default value (10 seconds),
         # which tells how much time Patroni will wait before checking the configuration
         # file again to reload it.

src/cluster.py

@@ -791,7 +791,7 @@ def remove_raft_member(self, member_ip: str) -> None:
         if not result.startswith("SUCCESS"):
             raise RemoveRaftMemberFailedError()
 
-    @retry(stop=stop_after_attempt(10), wait=wait_exponential(multiplier=1, min=2, max=10))
+    @retry(stop=stop_after_attempt(20), wait=wait_exponential(multiplier=1, min=2, max=10))
     def reload_patroni_configuration(self):
         """Reload Patroni configuration after it was changed."""
         requests.post(

tests/unit/test_charm.py

@@ -1338,6 +1338,7 @@ def test_update_config(harness):
         harness.update_relation_data(
             rel_id, harness.charm.unit.name, {"tls": ""}
         )  # Mock some data in the relation to test that it doesn't change.
+        _is_tls_enabled.return_value = False
         harness.charm.update_config()
         _handle_postgresql_restart_need.assert_not_called()
         assert "tls" not in harness.get_relation_data(rel_id, harness.charm.unit.name)