minor

app.py

@@ -149,6 +149,11 @@
         r = requests.patch(trial_url, data={"status": "error"},
                          headers = {"Authorization": "Token {}".format(API_TOKEN)})
         traceback.print_exc()
+
+        # Antoine: Removing this, it is too often causing the machines to stop. Not because
+        # the machines are failing, but because for instance the video is very long with a lot
+        # of people in it. We should not stop the machine for that. Originally the check was
+        # to catch a bug where the machine would hang, I have not seen this bug in a long time.
         # args_as_strings = [str(arg) for arg in e.args]
         # if len(args_as_strings) > 1 and 'pose detection timed out' in args_as_strings[1].lower():
         #     logging.info("Worker failed. Stopping machine.")

utils.py

@@ -1537,7 +1537,7 @@ def checkResourceUsage(stop_machine_and_email=True):
 
     if stop_machine_and_email and resourceUsage['disk_perc'] > 95:
 
-        message = "Disc is full on an OpenCap machine backend machine. It has been stopped. Data: " \
+        message = "Disc is full on an OpenCap backend machine. It has been stopped. Data: " \
                             + json.dumps(resourceUsage)
         sendStatusEmail(message=message)
 
@@ -1554,7 +1554,7 @@ def checkCudaTF():
         for gpu in gpus:
             print(f"GPU: {gpu.name}")
     else:
-        message = "Cuda check failed on an OpenCap machine backend machine. It has been stopped."
+        message = "Cuda check failed on an OpenCap backend machine. It has been stopped."
         sendStatusEmail(message=message)
         raise Exception("No GPU detected. Exiting.")