Add manifest arch, os, and compressed layers size fields

CHANGES/1767.feature

@@ -0,0 +1 @@
+Added `architecture`, `os`, and `compressed_layers_size` fields to Manifest.

pulp_container/app/management/commands/container-handle-image-data.py

@@ -1,3 +1,4 @@
+import json
 from json.decoder import JSONDecodeError
 
 from gettext import gettext as _
@@ -39,10 +40,22 @@ class Command(BaseCommand):
     def handle(self, *args, **options):
         manifests_updated_count = 0
 
-        manifests_v1 = Manifest.objects.filter(data__isnull=True, media_type=MEDIA_TYPE.MANIFEST_V1)
+        manifests_v1 = Manifest.objects.filter(
+            Q(media_type=MEDIA_TYPE.MANIFEST_V1),
+            Q(data__isnull=True)
+            | Q(architecture__isnull=True)
+            | Q(os__isnull=True)
+            | Q(compressed_layers_size__isnull=True),
+        )
         manifests_updated_count += self.update_manifests(manifests_v1)
 
-        manifests_v2 = Manifest.objects.filter(Q(data__isnull=True) | Q(annotations={}, labels={}))
+        manifests_v2 = Manifest.objects.filter(
+            Q(data__isnull=True)
+            | Q(annotations={}, labels={})
+            | Q(architecture__isnull=True)
+            | Q(os__isnull=True)
+            | Q(compressed_layers_size__isnull=True)
+        )
         manifests_v2 = manifests_v2.exclude(
             media_type__in=[MEDIA_TYPE.MANIFEST_LIST, MEDIA_TYPE.INDEX_OCI, MEDIA_TYPE.MANIFEST_V1]
         )
@@ -68,6 +81,17 @@ def handle(self, *args, **options):
     def update_manifests(self, manifests_qs):
         manifests_updated_count = 0
         manifests_to_update = []
+        fields_to_update = [
+            "annotations",
+            "labels",
+            "is_bootable",
+            "is_flatpak",
+            "data",
+            "os",
+            "architecture",
+            "compressed_layers_size",
+        ]
+
         for manifest in manifests_qs.iterator():
             # suppress non-existing/already migrated artifacts and corrupted JSON files
             with suppress(ObjectDoesNotExist, JSONDecodeError):
@@ -76,7 +100,6 @@ def update_manifests(self, manifests_qs):
                     manifests_to_update.append(manifest)
 
             if len(manifests_to_update) > 1000:
-                fields_to_update = ["annotations", "labels", "is_bootable", "is_flatpak", "data"]
                 manifests_qs.model.objects.bulk_update(
                     manifests_to_update,
                     fields_to_update,
@@ -85,7 +108,6 @@ def update_manifests(self, manifests_qs):
                 manifests_to_update.clear()
 
         if manifests_to_update:
-            fields_to_update = ["annotations", "labels", "is_bootable", "is_flatpak", "data"]
             manifests_qs.model.objects.bulk_update(
                 manifests_to_update,
                 fields_to_update,
@@ -103,8 +125,23 @@ def init_manifest(self, manifest):
             if not (manifest.annotations or manifest.labels):
                 manifest.init_metadata(manifest_data)
 
+            if self.needs_os_arch_size_update(manifest):
+                self.init_manifest_os_arch_size(manifest)
             manifest._artifacts.clear()
+            return True
 
+        elif self.needs_os_arch_size_update(manifest):
+            self.init_manifest_os_arch_size(manifest)
             return True
 
         return False
+
+    def needs_os_arch_size_update(self, manifest):
+        return manifest.media_type not in [MEDIA_TYPE.MANIFEST_LIST, MEDIA_TYPE.INDEX_OCI] and not (
+            manifest.architecture or manifest.os or manifest.compressed_layers_size
+        )
+
+    def init_manifest_os_arch_size(self, manifest):
+        manifest_data = json.loads(manifest.data)
+        manifest.init_architecture_and_os(manifest_data)
+        manifest.init_compressed_layers_size(manifest_data)

pulp_container/app/migrations/0042_update_manifest_and_manifestlistmanifest.py

@@ -0,0 +1,49 @@
+# Generated by Django 4.2.16 on 2024-10-02 12:04
+import warnings
+
+from django.db import migrations, models
+
+def print_warning_for_updating_manifest_fields(apps, schema_editor):
+    warnings.warn(
+        "Run 'pulpcore-manager container-handle-image-data' to update the manifests' "
+        "os, architecture, and compressed_layers_size fields."
+    )
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('container', '0041_add_pull_through_pull_permissions'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='manifest',
+            name='architecture',
+            field=models.TextField(null=True),
+        ),
+        migrations.AddField(
+            model_name='manifest',
+            name='compressed_layers_size',
+            field=models.IntegerField(null=True),
+        ),
+        migrations.AddField(
+            model_name='manifest',
+            name='os',
+            field=models.TextField(null=True),
+        ),
+        migrations.AlterField(
+            model_name='manifestlistmanifest',
+            name='architecture',
+            field=models.TextField(blank=True, default=''),
+        ),
+        migrations.AlterField(
+            model_name='manifestlistmanifest',
+            name='os',
+            field=models.TextField(blank=True, default=''),
+        ),
+        migrations.RunPython(
+            print_warning_for_updating_manifest_fields,
+            reverse_code=migrations.RunPython.noop,
+            elidable=True,
+        ),
+    ]

pulp_container/app/models.py

@@ -77,6 +77,11 @@ class Manifest(Content):
         labels (models.JSONField): Metadata stored inside the image configuration.
         is_bootable (models.BooleanField): Indicates whether the image is bootable or not.
         is_flatpak (models.BooleanField): Indicates whether the image is a flatpak package or not.
+        architecture (models.TextField): CPU architecture for which the binaries in the image are
+            designed to run.
+        os (models.TextField): Operating System which the image is built to run on.
+        compressed_layers_size (models.IntegerField): Sum of the sizes, in bytes, of all compressed
+            layers.
 
     Relations:
         blobs (models.ManyToManyField): Many-to-many relationship with Blob.
@@ -103,6 +108,9 @@ class Manifest(Content):
 
     annotations = models.JSONField(default=dict)
     labels = models.JSONField(default=dict)
+    architecture = models.TextField(null=True)
+    os = models.TextField(null=True)
+    compressed_layers_size = models.IntegerField(null=True)
 
     is_bootable = models.BooleanField(default=False)
     is_flatpak = models.BooleanField(default=False)
@@ -176,6 +184,32 @@ def init_manifest_nature(self):
         else:
             return False
 
+    def init_architecture_and_os(self, manifest_data):
+        # schema1 has the architecture/os definition in the Manifest (not in the ConfigBlob)
+        if manifest_data.get("architecture", None) or manifest_data.get("os", None):
+            self.architecture = manifest_data.get("architecture", None)
+            self.os = manifest_data.get("os", None)
+            return
+
+        manifest_config = manifest_data.get("config", None)
+        config_blob_sha256 = manifest_config.get("digest", None)
+        blob_artifact = Artifact.objects.get(sha256=config_blob_sha256.removeprefix("sha256:"))
+        config_blob, _ = get_content_data(blob_artifact)
+        self.architecture = config_blob.get("architecture", None)
+        self.os = config_blob.get("os", None)
+
+    def init_compressed_layers_size(self, manifest_data):
+        # manifestv2 schema1 has only blobSum definition for each layer
+        if manifest_data.get("fsLayers", None):
+            self.compressed_layers_size = 0
+            return
+
+        layers = manifest_data.get("layers")
+        compressed_size = 0
+        for layer in layers:
+            compressed_size += layer.get("size")
+        self.compressed_layers_size = compressed_size
+
     def is_bootable_image(self):
         if (
             self.annotations.get("containers.bootc") == "1"
@@ -222,8 +256,9 @@ class ManifestListManifest(models.Model):
         manifest_list (models.ForeignKey): Many-to-one relationship with ManifestList.
     """
 
-    architecture = models.TextField()
-    os = models.TextField()
+    # in oci-index spec, platform is an optional field
+    architecture = models.TextField(default="", blank=True)
+    os = models.TextField(default="", blank=True)
     os_version = models.TextField(default="", blank=True)
     os_features = models.TextField(default="", blank=True)
     features = models.TextField(default="", blank=True)

pulp_container/app/registry.py

@@ -454,10 +454,12 @@ async def init_pending_content(self, digest, manifest_data, media_type, raw_text
             config_blob=config_blob,
             data=raw_text_data,
         )
+        await sync_to_async(manifest.init_architecture_and_os)(manifest_data)
 
         # skip if media_type of schema1
         if media_type in (MEDIA_TYPE.MANIFEST_V2, MEDIA_TYPE.MANIFEST_OCI):
             await sync_to_async(manifest.init_metadata)(manifest_data=manifest_data)
+        await sync_to_async(manifest.init_compressed_layers_size)(manifest_data=manifest_data)
 
         try:
             await manifest.asave()

pulp_container/app/registry_api.py

@@ -1203,7 +1203,7 @@ def put(self, request, path, pk=None):
         if is_manifest_list:
             manifests = {}
             for manifest in content_data.get("manifests"):
-                manifests[manifest["digest"]] = manifest["platform"]
+                manifests[manifest["digest"]] = manifest.get("platform", None)
 
             digests = set(manifests.keys())
 
@@ -1217,17 +1217,17 @@ def put(self, request, path, pk=None):
 
             manifests_to_list = []
             for manifest in found_manifests:
-                platform = manifests[manifest.digest]
                 manifest_to_list = models.ManifestListManifest(
                     manifest_list=manifest,
                     image_manifest=manifest_list,
-                    architecture=platform["architecture"],
-                    os=platform["os"],
-                    features=platform.get("features", ""),
-                    variant=platform.get("variant", ""),
-                    os_version=platform.get("os.version", ""),
-                    os_features=platform.get("os.features", ""),
                 )
+                if platform := manifests[manifest.digest]:
+                    manifest_to_list.architecture = (platform["architecture"],)
+                    manifest_to_list.os = (platform["os"],)
+                    manifest_to_list.features = (platform.get("features", ""),)
+                    manifest_to_list.variant = (platform.get("variant", ""),)
+                    manifest_to_list.os_version = (platform.get("os.version", ""),)
+                    manifest_to_list.os_features = (platform.get("os.features", ""),)
                 manifests_to_list.append(manifest_to_list)
 
             models.ManifestListManifest.objects.bulk_create(
@@ -1295,6 +1295,7 @@ def put(self, request, path, pk=None):
             config_blob = found_config_blobs.first()
             manifest = self._init_manifest(manifest_digest, media_type, raw_text_data, config_blob)
             manifest.init_metadata(manifest_data=content_data)
+            manifest.init_compressed_layers_size(manifest_data=content_data)
 
             manifest = self._save_manifest(manifest)
 
@@ -1351,13 +1352,16 @@ def put(self, request, path, pk=None):
             return ManifestResponse(manifest, path, request, status=201)
 
     def _init_manifest(self, manifest_digest, media_type, raw_text_data, config_blob=None):
-        return models.Manifest(
+        manifest = models.Manifest(
             digest=manifest_digest,
             schema_version=2,
             media_type=media_type,
             config_blob=config_blob,
             data=raw_text_data,
         )
+        if config_blob:
+            manifest.init_architecture_and_os(json.loads(raw_text_data))
+        return manifest
 
     def _save_manifest(self, manifest):
         try:

pulp_container/app/serializers.py

@@ -103,6 +103,21 @@ class ManifestSerializer(NoArtifactContentSerializer):
         default=False,
         help_text=_("A boolean determining whether the image bundles a Flatpak application"),
     )
+    architecture = serializers.CharField(
+        help_text="The CPU architecture which the binaries in this image are built to run on.",
+        required=False,
+        default=None,
+    )
+    os = serializers.CharField(
+        help_text="The name of the operating system which the image is built to run on.",
+        required=False,
+        default=None,
+    )
+    compressed_layers_size = serializers.IntegerField(
+        help_text="Specifies the sum of the sizes, in bytes, of all compressed layers",
+        required=False,
+        default=None,
+    )
 
     class Meta:
         fields = NoArtifactContentSerializer.Meta.fields + (
@@ -116,6 +131,9 @@ class Meta:
             "labels",
             "is_bootable",
             "is_flatpak",
+            "architecture",
+            "os",
+            "compressed_layers_size",
         )
         model = models.Manifest
 

pulp_container/app/tasks/builder.py

@@ -86,11 +86,15 @@ def add_image_from_directory_to_repository(path, repository, tag):
 
         config_blob = get_or_create_blob(manifest_json["config"], manifest, path)
         manifest.config_blob = config_blob
-        manifest.save()
+        manifest.init_architecture_and_os(manifest_json)
 
         pks_to_add = []
+        compressed_size = 0
         for layer in manifest_json["layers"]:
+            compressed_size += layer.get("size")
             pks_to_add.append(get_or_create_blob(layer, manifest, path).pk)
+        manifest.compressed_layers_size = compressed_size
+        manifest.save()
 
         pks_to_add.extend([manifest.pk, tag.pk, config_blob.pk])
         new_repo_version.add_content(Content.objects.filter(pk__in=pks_to_add))