Fix constructed inventory host job summaries and relink orphaned summaries (#577)

The job_host_summaries API endpoint for hosts and groups in constructed
inventories was returning empty results because it queried the host FK
(which points to the source inventory host) instead of the
constructed_host FK. This matches the fix already present in the
serializer for recent_jobs.

Also adds a relink step after inventory sync that reconnects orphaned
JobHostSummary records (where the host FK became NULL after the host
was deleted by an overwrite sync) to the new host object by matching
on host_name, scoped to the same inventory to prevent cross-inventory
mis-linking.

Author: fernandorocagonzalez

awx/api/views/__init__.py

@@ -3711,10 +3711,20 @@ class BaseJobHostSummariesList(SubListAPIView):
 class HostJobHostSummariesList(BaseJobHostSummariesList):
     parent_model = models.Host
 
+    def get_sublist_queryset(self, parent):
+        if parent.inventory and parent.inventory.kind == 'constructed':
+            return parent.constructed_host_summaries
+        return super().get_sublist_queryset(parent)
+
 
 class GroupJobHostSummariesList(BaseJobHostSummariesList):
     parent_model = models.Group
 
+    def get_sublist_queryset(self, parent):
+        if parent.inventory and parent.inventory.kind == 'constructed':
+            return parent.constructed_host_summaries
+        return super().get_sublist_queryset(parent)
+
 
 class JobJobHostSummariesList(BaseJobHostSummariesList):
     parent_model = models.Job

awx/main/management/commands/inventory_import.py

@@ -23,6 +23,7 @@
 
 # AWX inventory imports
 from awx.main.models.inventory import Inventory, InventorySource, InventoryUpdate, Host
+from awx.main.models.jobs import JobHostSummary
 from awx.main.utils.mem_inventory import MemInventory, dict_to_mem_data
 from awx.main.utils.safe_yaml import sanitize_jinja
 
@@ -764,6 +765,26 @@ def _create_update_group_hosts(self):
         if settings.SQL_DEBUG:
             logger.warning('Group-host updates took %d queries for %d group-host relationships', len(connection.queries) - queries_before, group_host_count)
 
+    def _relink_orphaned_job_host_summaries(self):
+        host_map = dict(self.inventory.hosts.values_list('name', 'pk'))
+        if not host_map:
+            return
+
+        host_names = list(host_map.keys())
+        is_constructed = self.inventory.kind == 'constructed'
+        fk_field = 'constructed_host_id' if is_constructed else 'host_id'
+        null_filter = {'constructed_host__isnull': True} if is_constructed else {'host__isnull': True}
+
+        for offset in range(0, len(host_names), self._batch_size):
+            batch_names = host_names[offset : offset + self._batch_size]
+            orphaned = JobHostSummary.objects.filter(job__inventory=self.inventory, host_name__in=batch_names, **null_filter)
+            for host_name in batch_names:
+                host_pk = host_map.get(host_name)
+                if host_pk:
+                    updated = orphaned.filter(host_name=host_name).update(**{fk_field: host_pk})
+                    if updated:
+                        logger.info('Re-linked %d orphaned job host summaries for host "%s"', updated, host_name)
+
     def load_into_database(self):
         """
         Load inventory from in-memory groups to the database, overwriting or
@@ -784,6 +805,7 @@ def load_into_database(self):
         self._create_update_hosts(pk_mem_host_map)
         self._create_update_group_children()
         self._create_update_group_hosts()
+        self._relink_orphaned_job_host_summaries()
 
     def remote_tower_license_compare(self, local_license_type):
         # this requires https://github.com/ansible/ansible/pull/52747

awx/main/models/inventory.py

@@ -876,6 +876,12 @@ def job_host_summaries(self):
 
         return JobHostSummary.objects.filter(host__in=self.all_hosts)
 
+    @property
+    def constructed_host_summaries(self):
+        from awx.main.models.jobs import JobHostSummary
+
+        return JobHostSummary.objects.filter(constructed_host__in=self.all_hosts)
+
     @property
     def job_events(self):
         from awx.main.models.jobs import JobEvent

awx/main/tests/functional/commands/test_inventory_import.py

@@ -398,3 +398,138 @@ def test_tower_version_compare():
     with pytest.raises(PermissionDenied):
         cmd.remote_tower_license_compare('very_supported')
     cmd.remote_tower_license_compare('open')
+
+
+@pytest.mark.django_db
+@mock.patch.object(inventory_import.Command, 'set_logging_level', mock_logging)
+class TestRelinkOrphanedJobHostSummaries:
+    """After an overwrite sync deletes and recreates a host, orphaned
+    JobHostSummary records (host_id=NULL) should be re-linked to the
+    new host object by matching on host_name."""
+
+    def test_relink_after_host_recreated(self, inventory):
+        from awx.main.models import JobHostSummary, Job, Project, JobTemplate
+
+        inv_src = InventorySource.objects.create(inventory=inventory, source='ec2')
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=inventory, project=project)
+
+        data = {
+            '_meta': {'hostvars': {'server1': {}, 'server2': {}}},
+            'ungrouped': {'hosts': ['server1', 'server2']},
+        }
+        options = dict(overwrite=True)
+
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src.create_unified_job())
+        host1 = inventory.hosts.get(name='server1')
+
+        job = Job.objects.create(inventory=inventory, job_template=jt, status='successful')
+        JobHostSummary.objects.create(job=job, host=host1, host_name='server1', ok=1)
+
+        # Simulate host disappearing and reappearing (delete + reimport)
+        host1.delete()
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src.create_unified_job())
+
+        new_host = inventory.hosts.get(name='server1')
+        assert new_host.pk != host1.pk
+
+        summary = JobHostSummary.objects.get(job=job, host_name='server1')
+        assert summary.host_id == new_host.pk
+
+    def test_no_relink_when_host_still_linked(self, inventory):
+        from awx.main.models import JobHostSummary, Job, Project, JobTemplate
+
+        inv_src = InventorySource.objects.create(inventory=inventory, source='ec2')
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=inventory, project=project)
+
+        data = {
+            '_meta': {'hostvars': {'server1': {}}},
+            'ungrouped': {'hosts': ['server1']},
+        }
+        options = dict(overwrite=True)
+
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src.create_unified_job())
+        host1 = inventory.hosts.get(name='server1')
+
+        job = Job.objects.create(inventory=inventory, job_template=jt, status='successful')
+        JobHostSummary.objects.create(job=job, host=host1, host_name='server1', ok=1)
+
+        # Sync again without host disappearing - PK should be preserved
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src.create_unified_job())
+        same_host = inventory.hosts.get(name='server1')
+        assert same_host.pk == host1.pk
+
+        summary = JobHostSummary.objects.get(job=job, host_name='server1')
+        assert summary.host_id == host1.pk
+
+    def test_relink_constructed_inventory(self, organization):
+        from awx.main.models import JobHostSummary, Job, Project, JobTemplate
+
+        source_inv = Inventory.objects.create(name='source-inv', organization=organization)
+        constructed_inv = Inventory.objects.create(name='constructed-inv', kind='constructed', organization=organization)
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=constructed_inv, project=project)
+
+        source_host = Host.objects.create(name='server1', inventory=source_inv)
+        constructed_host = Host.objects.create(
+            name='server1', inventory=constructed_inv, instance_id=str(source_host.pk)
+        )
+
+        job = Job.objects.create(inventory=constructed_inv, job_template=jt, status='successful')
+        JobHostSummary.objects.create(
+            job=job, host=source_host, constructed_host=constructed_host,
+            host_name='server1', ok=1
+        )
+
+        old_constructed_pk = constructed_host.pk
+        constructed_host.delete()
+
+        # Recreate constructed host (simulates constructed inventory re-sync)
+        new_constructed_host = Host.objects.create(
+            name='server1', inventory=constructed_inv, instance_id=str(source_host.pk)
+        )
+
+        inv_src = InventorySource.objects.create(inventory=constructed_inv, source='constructed')
+        data = {
+            '_meta': {'hostvars': {'server1': {}}},
+            'ungrouped': {'hosts': ['server1']},
+        }
+        inventory_import.Command().perform_update(dict(overwrite=True), data, inv_src.create_unified_job())
+
+        summary = JobHostSummary.objects.get(job=job, host_name='server1')
+        assert summary.constructed_host_id is not None
+        assert summary.constructed_host_id != old_constructed_pk
+
+    def test_relink_does_not_cross_inventories(self, organization):
+        from awx.main.models import JobHostSummary, Job, Project, JobTemplate
+
+        inv_a = Inventory.objects.create(name='inv-a', organization=organization)
+        inv_b = Inventory.objects.create(name='inv-b', organization=organization)
+        inv_src_a = InventorySource.objects.create(inventory=inv_a, source='ec2')
+        inv_src_b = InventorySource.objects.create(inventory=inv_b, source='ec2')
+        project = Project.objects.create(name='test-proj')
+
+        data = {
+            '_meta': {'hostvars': {'server1': {}}},
+            'ungrouped': {'hosts': ['server1']},
+        }
+        options = dict(overwrite=True)
+
+        # Create host in both inventories
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src_a.create_unified_job())
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src_b.create_unified_job())
+
+        host_b = inv_b.hosts.get(name='server1')
+        jt_b = JobTemplate.objects.create(name='test-jt-b', inventory=inv_b, project=project)
+        job_b = Job.objects.create(inventory=inv_b, job_template=jt_b, status='successful')
+        JobHostSummary.objects.create(job=job_b, host=host_b, host_name='server1', ok=1)
+
+        # Delete host from inv_b, orphaning the summary
+        host_b.delete()
+
+        # Sync inv_a: should NOT re-link inv_b's orphaned summary
+        inventory_import.Command().perform_update(options.copy(), data.copy(), inv_src_a.create_unified_job())
+
+        summary = JobHostSummary.objects.get(job=job_b, host_name='server1')
+        assert summary.host_id is None, "Summary from inv_b should not be re-linked to inv_a's host"

awx/main/tests/functional/models/test_host_summary_fields.py

@@ -2,8 +2,9 @@
 
 from django.utils.timezone import now
 
-from awx.main.models import Job, JobEvent, JobTemplate, Inventory, Host, JobHostSummary, Project
+from awx.main.models import Job, JobEvent, JobTemplate, Inventory, Host, Group, JobHostSummary, Project
 from awx.api.serializers import HostSerializer
+from awx.api.versioning import reverse
 
 
 @pytest.mark.django_db
@@ -109,3 +110,91 @@ def test_no_summary_fields_without_job(self):
 
         assert 'last_job' not in d
         assert 'last_job_host_summary' not in d
+
+
+@pytest.mark.django_db
+class TestConstructedHostJobSummariesAPI:
+    """The job_host_summaries endpoint for hosts in a constructed inventory
+    should use the constructed_host FK, not the regular host FK."""
+
+    def test_constructed_host_summaries_returned(self, get, admin, organization):
+        source_inv = Inventory.objects.create(name='source', organization=organization)
+        constructed_inv = Inventory.objects.create(name='constructed', kind='constructed', organization=organization)
+
+        source_host = Host.objects.create(name='server1', inventory=source_inv)
+        constructed_host = Host.objects.create(name='server1', inventory=constructed_inv, instance_id=str(source_host.pk))
+
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=constructed_inv, project=project)
+        job = Job.objects.create(inventory=constructed_inv, job_template=jt, status='successful')
+
+        JobHostSummary.objects.create(
+            job=job, host=source_host, constructed_host=constructed_host,
+            host_name='server1', ok=1
+        )
+
+        url = reverse('api:host_job_host_summaries_list', kwargs={'pk': constructed_host.pk})
+        resp = get(url, user=admin, expect=200)
+        assert resp.data['count'] == 1
+        assert resp.data['results'][0]['host_name'] == 'server1'
+
+    def test_regular_host_summaries_still_work(self, get, admin, organization):
+        inv = Inventory.objects.create(name='regular', organization=organization)
+        host = Host.objects.create(name='server1', inventory=inv)
+
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=inv, project=project)
+        job = Job.objects.create(inventory=inv, job_template=jt, status='successful')
+
+        JobHostSummary.objects.create(job=job, host=host, host_name='server1', ok=1)
+
+        url = reverse('api:host_job_host_summaries_list', kwargs={'pk': host.pk})
+        resp = get(url, user=admin, expect=200)
+        assert resp.data['count'] == 1
+
+    def test_constructed_host_no_false_positives(self, get, admin, organization):
+        source_inv = Inventory.objects.create(name='source', organization=organization)
+        constructed_inv = Inventory.objects.create(name='constructed', kind='constructed', organization=organization)
+
+        source_host = Host.objects.create(name='server1', inventory=source_inv)
+        constructed_host = Host.objects.create(name='server1', inventory=constructed_inv, instance_id=str(source_host.pk))
+
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=source_inv, project=project)
+        job = Job.objects.create(inventory=source_inv, job_template=jt, status='successful')
+
+        # Summary linked to source host only, not the constructed one
+        JobHostSummary.objects.create(job=job, host=source_host, host_name='server1', ok=1)
+
+        url = reverse('api:host_job_host_summaries_list', kwargs={'pk': constructed_host.pk})
+        resp = get(url, user=admin, expect=200)
+        assert resp.data['count'] == 0
+
+
+@pytest.mark.django_db
+class TestConstructedGroupJobSummariesAPI:
+    """The job_host_summaries endpoint for groups in a constructed inventory
+    should use the constructed_host FK."""
+
+    def test_constructed_group_summaries_returned(self, get, admin, organization):
+        source_inv = Inventory.objects.create(name='source', organization=organization)
+        constructed_inv = Inventory.objects.create(name='constructed', kind='constructed', organization=organization)
+
+        source_host = Host.objects.create(name='server1', inventory=source_inv)
+        constructed_host = Host.objects.create(name='server1', inventory=constructed_inv, instance_id=str(source_host.pk))
+
+        group = Group.objects.create(name='webservers', inventory=constructed_inv)
+        group.hosts.add(constructed_host)
+
+        project = Project.objects.create(name='test-proj')
+        jt = JobTemplate.objects.create(name='test-jt', inventory=constructed_inv, project=project)
+        job = Job.objects.create(inventory=constructed_inv, job_template=jt, status='successful')
+
+        JobHostSummary.objects.create(
+            job=job, host=source_host, constructed_host=constructed_host,
+            host_name='server1', ok=1
+        )
+
+        url = reverse('api:group_job_host_summaries_list', kwargs={'pk': group.pk})
+        resp = get(url, user=admin, expect=200)
+        assert resp.data['count'] == 1