fix

Author: hanwen-cluster

src/slurm_plugin/fleet_manager.py

@@ -200,8 +200,8 @@ def launch_ec2_instances(self, count, job_id=None):
                 )
                 logger.debug("Launched instances information: %s", assigned_nodes.get("Instances"))
                 running_nodes_file_path = "/etc/parallelcluster/slurm_plugin/running_nodes"
-                with open(running_nodes_file_path, "a+") as f:
-                    f.writelines(instance_ids)
+                with open(running_nodes_file_path, "a") as f:
+                    f.write('\n'.join(instance_ids)+'\n')
 
         return [EC2Instance.from_describe_instance_data(instance_info) for instance_info in assigned_nodes["Instances"]]
 

src/slurm_plugin/instance_manager.py

@@ -266,8 +266,8 @@ def get_cluster_instances(self, include_head_node=False):
         """
         running_instances_from_file = set()
         running_instances_file_path = "/etc/parallelcluster/slurm_plugin/running_nodes"
-        with open(running_instances_file_path, "a+") as f:
-            running_instances_from_file.update(set(f.readlines()))
+        with open(running_instances_file_path, "r") as f:
+            running_instances_from_file.update(set([line.strip() for line in f.readlines() if line.strip()]))
 
         logger.info("hellllllllllllll1")
         logger.info(running_instances_from_file)
@@ -300,7 +300,7 @@ def get_cluster_instances(self, include_head_node=False):
             logger.info("hellllllllllllll2")
             logger.info(running_instances_from_file)
             logger.info(non_existing_instances)
-            f.writelines(list(running_instances_from_file - non_existing_instances))
+            f.write('\n'.join(list(running_instances_from_file - non_existing_instances))+'\n')
         return instances