Revise start_fact_cache and finish_fact_cache to use JSON file (#15970)

* Revise start_fact_cache and finish_fact_cache to use JSON file with host list inside it

* Revise artifacts path to be relative to the job private_data_dir

* Update calls to start_fact_cache and finish_fact_cache to agree with new reference to artifacts_dir

* Prevents unnecessary updates to ansible_facts_modified, fixing timestamp-related test failures.

awx/main/tasks/facts.py

@@ -22,27 +22,29 @@ system_tracking_logger = logging.getLogger('awx.analytics.system_tracking')
 
 
 @log_excess_runtime(logger, debug_cutoff=0.01, msg='Inventory {inventory_id} host facts prepared for {written_ct} hosts, took {delta:.3f} s', add_log_data=True)
-def start_fact_cache(hosts, destination, log_data, timeout=None, inventory_id=None):
+def start_fact_cache(hosts, artifacts_dir, timeout=None, inventory_id=None, log_data=None):
+    log_data = log_data or {}
     log_data['inventory_id'] = inventory_id
     log_data['written_ct'] = 0
-    hosts_cached = list()
-    try:
-        os.makedirs(destination, mode=0o700)
-    except FileExistsError:
-        pass
+    hosts_cached = []
+
+    # Create the fact_cache directory inside artifacts_dir
+    fact_cache_dir = os.path.join(artifacts_dir, 'fact_cache')
+    os.makedirs(fact_cache_dir, mode=0o700, exist_ok=True)
 
     if timeout is None:
         timeout = settings.ANSIBLE_FACT_CACHE_TIMEOUT
 
-    last_filepath_written = None
+    last_write_time = None
+
     for host in hosts:
-        hosts_cached.append(host)
+        hosts_cached.append(host.name)
         if not host.ansible_facts_modified or (timeout and host.ansible_facts_modified < now() - datetime.timedelta(seconds=timeout)):
             continue  # facts are expired - do not write them
 
-        filepath = os.sep.join(map(str, [destination, host.name]))
-        if not os.path.realpath(filepath).startswith(destination):
-            system_tracking_logger.error('facts for host {} could not be cached'.format(smart_str(host.name)))
+        filepath = os.path.join(fact_cache_dir, host.name)
+        if not os.path.realpath(filepath).startswith(fact_cache_dir):
+            logger.error(f'facts for host {smart_str(host.name)} could not be cached')
             continue
 
         try:
@@ -50,15 +52,21 @@ def start_fact_cache(hosts, destination, log_data, timeout=None, inventory_id=No
                 os.chmod(f.name, 0o600)
                 json.dump(host.ansible_facts, f)
                 log_data['written_ct'] += 1
-                last_filepath_written = filepath
+                last_write_time = os.path.getmtime(filepath)
         except IOError:
-            system_tracking_logger.error('facts for host {} could not be cached'.format(smart_str(host.name)))
+            logger.error(f'facts for host {smart_str(host.name)} could not be cached')
             continue
 
-    if last_filepath_written:
-        return os.path.getmtime(last_filepath_written), hosts_cached
-
-    return None, hosts_cached
+    # Write summary file directly to the artifacts_dir
+    if inventory_id is not None:
+        summary_file = os.path.join(artifacts_dir, 'host_cache_summary.json')
+        summary_data = {
+            'last_write_time': last_write_time,
+            'hosts_cached': hosts_cached,
+            'written_ct': log_data['written_ct'],
+        }
+        with open(summary_file, 'w', encoding='utf-8') as f:
+            json.dump(summary_data, f, indent=2)
 
 
 @log_excess_runtime(
@@ -67,34 +75,54 @@ def start_fact_cache(hosts, destination, log_data, timeout=None, inventory_id=No
     msg='Inventory {inventory_id} host facts: updated {updated_ct}, cleared {cleared_ct}, unchanged {unmodified_ct}, took {delta:.3f} s',
     add_log_data=True,
 )
-def finish_fact_cache(hosts_cached, destination, facts_write_time, log_data, job_id=None, inventory_id=None):
+def finish_fact_cache(artifacts_dir, job_id=None, inventory_id=None, log_data=None):
+    log_data = log_data or {}
     log_data['inventory_id'] = inventory_id
     log_data['updated_ct'] = 0
     log_data['unmodified_ct'] = 0
     log_data['cleared_ct'] = 0
+    # The summary file is directly inside the artifacts dir
+    summary_path = os.path.join(artifacts_dir, 'host_cache_summary.json')
+    if not os.path.exists(summary_path):
+        logger.error(f'Missing summary file at {summary_path}')
+        return
 
-    hosts_cached = sorted((h for h in hosts_cached if h.id is not None), key=lambda h: h.id)
+    try:
+        with open(summary_path, 'r', encoding='utf-8') as f:
+            summary = json.load(f)
+        facts_write_time = os.path.getmtime(summary_path)  # After successful read
+    except (json.JSONDecodeError, OSError) as e:
+        logger.error(f'Error reading summary file at {summary_path}: {e}')
+        return
 
+    host_names = summary.get('hosts_cached', [])
+    hosts_cached = Host.objects.filter(name__in=host_names).order_by('id').iterator()
+    # Path where individual fact files were written
+    fact_cache_dir = os.path.join(artifacts_dir, 'fact_cache')
     hosts_to_update = []
+
     for host in hosts_cached:
-        filepath = os.sep.join(map(str, [destination, host.name]))
-        if not os.path.realpath(filepath).startswith(destination):
-            system_tracking_logger.error('facts for host {} could not be cached'.format(smart_str(host.name)))
+        filepath = os.path.join(fact_cache_dir, host.name)
+        if not os.path.realpath(filepath).startswith(fact_cache_dir):
+            logger.error(f'Invalid path for facts file: {filepath}')
             continue
+
         if os.path.exists(filepath):
             # If the file changed since we wrote the last facts file, pre-playbook run...
             modified = os.path.getmtime(filepath)
-            if (not facts_write_time) or modified > facts_write_time:
-                with codecs.open(filepath, 'r', encoding='utf-8') as f:
-                    try:
+            if not facts_write_time or modified >= facts_write_time:
+                try:
+                    with codecs.open(filepath, 'r', encoding='utf-8') as f:
                         ansible_facts = json.load(f)
-                    except ValueError:
-                        continue
+                except ValueError:
+                    continue
+
+                if ansible_facts != host.ansible_facts:
                     host.ansible_facts = ansible_facts
                     host.ansible_facts_modified = now()
                     hosts_to_update.append(host)
-                    system_tracking_logger.info(
-                        'New fact for inventory {} host {}'.format(smart_str(host.inventory.name), smart_str(host.name)),
+                    logger.info(
+                        f'New fact for inventory {smart_str(host.inventory.name)} host {smart_str(host.name)}',
                         extra=dict(
                             inventory_id=host.inventory.id,
                             host_name=host.name,
@@ -104,6 +132,8 @@ def finish_fact_cache(hosts_cached, destination, facts_write_time, log_data, job
                         ),
                     )
                     log_data['updated_ct'] += 1
+                else:
+                    log_data['unmodified_ct'] += 1
             else:
                 log_data['unmodified_ct'] += 1
         else:
@@ -112,9 +142,11 @@ def finish_fact_cache(hosts_cached, destination, facts_write_time, log_data, job
             host.ansible_facts = {}
             host.ansible_facts_modified = now()
             hosts_to_update.append(host)
-            system_tracking_logger.info('Facts cleared for inventory {} host {}'.format(smart_str(host.inventory.name), smart_str(host.name)))
+            logger.info(f'Facts cleared for inventory {smart_str(host.inventory.name)} host {smart_str(host.name)}')
             log_data['cleared_ct'] += 1
-        if len(hosts_to_update) > 100:
+
+        if len(hosts_to_update) >= 100:
             bulk_update_sorted_by_id(Host, hosts_to_update, fields=['ansible_facts', 'ansible_facts_modified'])
             hosts_to_update = []
+
     bulk_update_sorted_by_id(Host, hosts_to_update, fields=['ansible_facts', 'ansible_facts_modified'])