fix k8 nightly release tests failures

runhouse/resources/hardware/on_demand_cluster.py

@@ -450,6 +450,12 @@ def _populate_connection_from_status_dict(self, cluster_dict: Dict[str, Any]):
                 ssh_values = backend_utils.ssh_credential_from_yaml(
                     yaml_path, ssh_user=handle.ssh_user
                 )
+                # removing unnecessary apostrophes form ssh_proxy_command, if such exist.
+                # meaning if ssh_proxy_command contains "...ProxyCommand='...' ", we are setting it to "...ProxyCommand=... "
+                if ssh_values["ssh_proxy_command"]:
+                    ssh_values["ssh_proxy_command"] = ssh_values.get(
+                        "ssh_proxy_command"
+                    ).replace("'", "")
                 if not self.creds_values or not self.ssh_properties:
                     self._setup_creds(ssh_values)
 
@@ -526,6 +532,9 @@ def _populate_connection_from_status_dict(self, cluster_dict: Dict[str, Any]):
                     _, current_context = kubernetes.config.list_kube_config_contexts()
                     self.compute_properties["kube_context"] = current_context["name"]
 
+                self._kube_namespace = self.compute_properties.get("kube_namespace")
+                self._kube_context = self.compute_properties.get("kube_context")
+
     def _update_from_sky_status(self, dryrun: bool = False):
         # Try to get the cluster status from SkyDB
         if self.is_shared:

tests/fixtures/on_demand_cluster_fixtures.py

@@ -186,7 +186,7 @@ def ondemand_k8s_cluster(request, test_rns_folder):
 
     # Note: Cannot specify both `instance_type` and any of `memory`, `disk_size`, `num_cpus`, or `accelerators`
     args = {
-        "name": "k8s-cpu",
+        "name": f"{test_rns_folder}-k8s-cpu",
         "provider": "kubernetes",
         "instance_type": "CPU:1",
     }