From f2ae68f30207da413655412d1146e6f7328c2d9d Mon Sep 17 00:00:00 2001
From: Alan Rominger <arominge@redhat.com>
Date: Tue, 10 Dec 2024 15:23:54 -0500
Subject: [PATCH] Fix project cache identifiers for new updates (#6762)

Finish test and discover viable solution

Add comment on related task code
---
 awx/main/models/projects.py                   | 26 ++++++++-
 awx/main/tasks/jobs.py                        |  1 +
 .../live/tests/projects/test_requirements.py  | 56 +++++++++++++++++++
 3 files changed, 81 insertions(+), 2 deletions(-)
 create mode 100644 awx/main/tests/live/tests/projects/test_requirements.py

diff --git a/awx/main/models/projects.py b/awx/main/models/projects.py
index 8c8fcd52ba..d0b22830ee 100644
--- a/awx/main/models/projects.py
+++ b/awx/main/models/projects.py
@@ -5,6 +5,8 @@
 import datetime
 import os
 import urllib.parse as urlparse
+from uuid import uuid4
+import logging
 
 # Django
 from django.conf import settings
@@ -39,6 +41,8 @@ from awx.main.models.rbac import (
     ROLE_SINGLETON_SYSTEM_AUDITOR,
 )
 
+logger = logging.getLogger('awx.main.models.projects')
+
 __all__ = ['Project', 'ProjectUpdate']
 
 
@@ -447,7 +451,25 @@ class Project(UnifiedJobTemplate, ProjectOptions, ResourceMixin, CustomVirtualEn
 
     @property
     def cache_id(self):
-        return str(self.last_job_id)
+        """This gives the folder name where collections and roles will be saved to so it does not re-download
+
+        Normally we want this to track with the last update, because every update should pull new content.
+        This does not count sync jobs, but sync jobs do not update last_job or current_job anyway.
+        If cleanup_jobs deletes the last jobs, then we can fallback to using any given heuristic related
+        to the last job ran.
+        """
+        if self.current_job_id:
+            return str(self.current_job_id)
+        elif self.last_job_id:
+            return str(self.last_job_id)
+        elif self.last_job_run:
+            return self.last_job_run.isoformat()
+        else:
+            logger.warning(f'No info about last update for project {self.id}, content cache may misbehave')
+            if self.modified:
+                return self.modified.isoformat()
+            else:
+                return str(uuid4())
 
     @property
     def notification_templates(self):
@@ -618,7 +640,7 @@ class ProjectUpdate(UnifiedJob, ProjectOptions, JobNotificationMixin, TaskManage
     @property
     def cache_id(self):
         if self.branch_override or self.job_type == 'check' or (not self.project):
-            return str(self.id)
+            return str(self.id)  # causes it to not use the cache, basically
         return self.project.cache_id
 
     def result_stdout_raw_limited(self, start_line=0, end_line=None, redact_sensitive=True):
diff --git a/awx/main/tasks/jobs.py b/awx/main/tasks/jobs.py
index c6cfc6a180..509ebc7e98 100644
--- a/awx/main/tasks/jobs.py
+++ b/awx/main/tasks/jobs.py
@@ -698,6 +698,7 @@ class SourceControlMixin(BaseTask):
             logger.debug(f'Project not available locally, {self.instance.id} will sync with remote')
             sync_needs.append(source_update_tag)
 
+        # Determine whether or not this project sync needs to populate the cache for Ansible content, roles and collections
         has_cache = os.path.exists(os.path.join(project.get_cache_path(), project.cache_id))
         # Galaxy requirements are not supported for manual projects
         if project.scm_type and ((not has_cache) or branch_override):
diff --git a/awx/main/tests/live/tests/projects/test_requirements.py b/awx/main/tests/live/tests/projects/test_requirements.py
new file mode 100644
index 0000000000..c82ccbec80
--- /dev/null
+++ b/awx/main/tests/live/tests/projects/test_requirements.py
@@ -0,0 +1,56 @@
+import os
+import time
+
+import pytest
+
+from django.conf import settings
+
+from awx.main.tests.live.tests.conftest import wait_for_job
+
+from awx.main.models import Project, SystemJobTemplate
+
+
+@pytest.fixture(scope='session')
+def project_with_requirements(default_org):
+    project, _ = Project.objects.get_or_create(
+        name='project-with-requirements',
+        scm_url='https://github.com/ansible/test-playbooks.git',
+        scm_branch="with_requirements",
+        scm_type='git',
+        organization=default_org,
+    )
+    start = time.time()
+    while time.time() - start < 3.0:
+        if project.current_job or project.last_job or project.last_job_run:
+            break
+    assert project.current_job or project.last_job or project.last_job_run, f'Project never updated id={project.id}'
+    update = project.current_job or project.last_job
+    if update:
+        wait_for_job(update)
+    return project
+
+
+def project_cache_is_populated(project):
+    proj_cache = os.path.join(project.get_cache_path(), project.cache_id)
+    return os.path.exists(proj_cache)
+
+
+def test_cache_is_populated_after_cleanup_job(project_with_requirements):
+    assert project_with_requirements.cache_id is not None  # already updated, should be something
+    cache_path = os.path.join(settings.PROJECTS_ROOT, '.__awx_cache')
+    assert os.path.exists(cache_path)
+
+    assert project_cache_is_populated(project_with_requirements)
+
+    cleanup_sjt = SystemJobTemplate.objects.get(name='Cleanup Job Details')
+    cleanup_job = cleanup_sjt.create_unified_job(extra_vars={'days': 0})
+    cleanup_job.signal_start()
+    wait_for_job(cleanup_job)
+
+    project_with_requirements.refresh_from_db()
+    assert project_with_requirements.cache_id is not None
+    update = project_with_requirements.update()
+    wait_for_job(update)
+
+    # Now, we still have a populated cache
+    assert project_cache_is_populated(project_with_requirements)