diff --git a/libs/wlgen/wlgen/rta.py b/libs/wlgen/wlgen/rta.py
index 813b30703b51982ba353ad63278cd8d2e0f58f6b..b26328b3781dae2a76f0a9780e49f9d8d3a78ade 100644
--- a/libs/wlgen/wlgen/rta.py
+++ b/libs/wlgen/wlgen/rta.py
@@ -222,8 +222,6 @@ class RTA(Workload):
 
         rtapp_conf = self.params['custom']
 
-        self._log.info('Loading custom configuration:')
-        self._log.info('   %s', rtapp_conf)
         self.json = '{0:s}_{1:02d}.json'.format(self.name, self.exc_id)
         ofile = open(self.json, 'w')
 
@@ -252,6 +250,8 @@ class RTA(Workload):
                                  "a filename or an embedded dictionary")
         else:
             # We assume we are being passed a filename instead of a dict
+            self._log.info('Loading custom configuration:')
+            self._log.info('   %s', rtapp_conf)
             ifile = open(rtapp_conf, 'r')
 
         for line in ifile:
diff --git a/tests/eas/load_tracking.py b/tests/eas/load_tracking.py
index c8a927e71a0a3bf161b3439387c94cdc35d5f165..fdc714e17922c5fe7781bae044f7c1bdc7e9b5d9 100644
--- a/tests/eas/load_tracking.py
+++ b/tests/eas/load_tracking.py
@@ -557,6 +557,9 @@ class _CPUMigrationBase(LisaTest):
     # Allowed error margin
     allowed_util_margin = UTIL_SCALE * 0.02
 
+    # Dictionary that contains the description of the tasks
+    tasks_desc = {}
+
     @classmethod
     def setUpClass(cls, *args, **kwargs):
         super(_CPUMigrationBase, cls).runExperiments(*args, **kwargs)
@@ -581,12 +584,13 @@ class _CPUMigrationBase(LisaTest):
         rtapp.conf(kind='profile',
             params={'task{}'.format(i): task.get() for i, task in enumerate(tasks)},
             run_dir=Executor.get_run_dir(test_env.target))
+        cls.tasks_desc = rtapp.rta_profile['tasks']
         return {
             'migration': {
                 'type' : 'rt-app',
                 'conf' : {
                     'class' : 'custom',
-                    'json' : rtapp.json,
+                    'json' : rtapp.rta_profile,
                     'prefix' : 'mig_test',
                 },
             }
@@ -743,6 +747,7 @@ class _CPUMigrationBase(LisaTest):
 
     def _test_util_per_cpu(self, experiment, tasks):
         trace = self.get_trace(experiment)
+
         if not trace.hasEvents('sched_switch'):
             raise ValueError('No sched_switch events. '
                              'Does the kernel support them?')
@@ -750,31 +755,29 @@ class _CPUMigrationBase(LisaTest):
             raise ValueError('No sched_load_cfs_rq events. '
                              'Does the kernel support them?')
         cpus = set()
-        # Load the JSON tasks description
-        tasks_desc = json.load(open(experiment.wload.params['custom']),
-                               object_pairs_hook=OrderedDict)['tasks']
-        sw_df = trace.data_frame.trace_event('sched_switch')
+
         # Filter the event related to the tasks
-        sw_df = sw_df[sw_df.next_comm.isin(tasks_desc.keys())]
+        sw_df = trace.data_frame.trace_event('sched_switch')
+        sw_df = sw_df[sw_df.next_comm.isin(self.tasks_desc.keys())]
 
         util_df = trace.data_frame.trace_event('sched_load_cfs_rq')
-        phases = self._get_phases_names(self._get_one_task(tasks_desc, 0))
+        phases = self._get_phases_names(self._get_one_task(self.tasks_desc, 0))
 
         # Compute the interval where the signal is stable for the phases
         window = self._get_stable_window(sw_df,
-                                         self._get_one_task(tasks_desc, 0))
+                                         self._get_one_task(self.tasks_desc, 0))
 
         msg = 'Saw util {} on cpu {}, expected {} during phase {}'
         for phase in phases:
             # Get all the cpus where tasks are running during this phase
-            for task in tasks_desc.iteritems():
+            for task in self.tasks_desc.iteritems():
                 cpus.update(self._get_cpus(task[1], phase))
 
             # Get the mean utilization per CPU
             util_mean = self._get_util_mean(window[phase][0], window[phase][1],
                                             util_df, cpus)
             # Get the expected utilization per CPU
-            expected = self._get_util_expected(tasks_desc, cpus, phase)
+            expected = self._get_util_expected(self.tasks_desc, cpus, phase)
 
             # Check that the expected utilization value and the measured one
             # match for each CPU.