[CoreEngine] fixed the issue when the product-id file is limited to access due to the permission.

Author: fedml-alex

devops/k8s/README_MODEL_SERVING.md

@@ -101,7 +101,7 @@ Moreover, on GCP k8s cluster, you should set up your GPU nodes based on the foll
 After you have installed FedML model serving packages, you may run the helm upgrade commands to modify parameters.
 
 e.g.
-```helm upgrade --set "autoscaling.enabled=true" --set replicaCount=$InstanceNumber fedml-model-premise-master fedml-model-premise-master-0.7.397.tgz -n $YourNameSpace```
+```helm upgrade --set "autoscaling.enabled=true" --set replicaCount=$InstanceNumber fedml-model-premise-master fedml-model-premise-master-latest.tgz -n $YourNameSpace```
 
 ### 6). Config your CNAME record in your DNS provider (Godaddy, wordpress, AWS Route 53...)
 #### (a). Find the Kubernetes nginx ingress named 'fedml-model-inference-gateway' in your Kubernetes cluster.
@@ -150,7 +150,7 @@ Pull remote model(ModelOps) to local model repository:
 1. Q: Supports automatically scale?
 A: Yes. Call CLI `helm upgrade`. For example, you can do upgrade by using the following CLI:
 
-```helm upgrade --set "autoscaling.enabled=true" --set replicaCount=$InstanceNumber fedml-model-premise-master fedml-model-premise-master-0.7.397.tgz -n $YourNameSpace```
+```helm upgrade --set "autoscaling.enabled=true" --set replicaCount=$InstanceNumber fedml-model-premise-master fedml-model-premise-master-latest.tgz -n $YourNameSpace```
 
 
 2. Q: Does the inference endpoint supports private IP? \

python/fedml/__init__.py

@@ -23,7 +23,7 @@
 _global_training_type = None
 _global_comm_backend = None
 
-__version__ = "0.8.2a3"
+__version__ = "0.8.2a4"
 
 
 def init(args=None):

python/fedml/cli/comm_utils/sys_utils.py

@@ -474,10 +474,14 @@ def get_device_id_in_docker():
 
     if os.path.exists(docker_env_file) or os.path.exists(cgroup_file):
         if os.path.exists(product_uuid_file):
-            with open(product_uuid_file, 'r') as f:
-                device_id = f.readline().rstrip("\n").strip(" ")
-                if device_id == "":
-                    device_id = str(uuid.uuid4())
+            try:
+                with open(product_uuid_file, 'r') as f:
+                    device_id = f.readline().rstrip("\n").strip(" ")
+                    if device_id == "":
+                        device_id = str(uuid.uuid4())
+                    return f"{device_id}-docker"
+            except Exception as e:
+                device_id = str(uuid.uuid4())
                 return f"{device_id}-docker"
     return None
 

python/fedml/core/mlops/mlops_runtime_log.py

@@ -119,7 +119,7 @@ def log_ntp_time(sec, what):
                 ntp_time_seconds = time.time()
             ntp_time = datetime.datetime.fromtimestamp(ntp_time_seconds)
             return ntp_time.timetuple()
-        # logging.Formatter.converter = log_ntp_time
+        logging.Formatter.converter = log_ntp_time
         stdout_handle = logging.StreamHandler()
         stdout_handle.setFormatter(format_str)
         if show_stdout_log:

python/setup.py

@@ -88,7 +88,7 @@ def finalize_options(self):
 
 setup(
     name="fedml",
-    version="0.8.2a3",
+    version="0.8.2a4",
     author="FedML Team",
     author_email="ch@fedml.ai",
     description="A research and production integrated edge-cloud library for "