[Fixes #14381] Relax Remote Service URL uniqueness

geonode/geoserver/signals.py

@@ -34,8 +34,8 @@
 from geonode.layers.models import Dataset
 from geonode.services.enumerations import CASCADED
 
-from . import BACKEND_PACKAGE
-from .tasks import geoserver_cascading_delete, geoserver_post_save_datasets
+from geonode.geoserver import BACKEND_PACKAGE
+from geonode.geoserver.tasks import geoserver_cascading_delete, geoserver_post_save_datasets
 
 logger = logging.getLogger("geonode.geoserver.signals")
 
@@ -104,14 +104,22 @@ def geoserver_pre_save_maplayer(instance, sender, **kwargs):
     # Set dataset
     if instance.dataset is None:
         dataset_queryset = Dataset.objects.filter(Q(alternate=instance.name) | Q(name=instance.name))
-        if instance.local and instance.store:
+        if instance.store:
             dataset_queryset = dataset_queryset.filter(store=instance.store)
         elif instance.ows_url:
             dataset_queryset = dataset_queryset.filter(remote_service__base_url=instance.ows_url)
         try:
             instance.dataset = dataset_queryset.get()
-        except (Dataset.DoesNotExist, Dataset.MultipleObjectsReturned):
+        except Dataset.DoesNotExist:
             pass
+        except Dataset.MultipleObjectsReturned:
+            logger.error(
+                "Ambiguous dataset lookup for map layer '%s' (store=%s, ows_url=%s)",
+                instance.name,
+                instance.store,
+                instance.ows_url,
+            )
+            raise
 
 
 @deprecated(version="3.2.1", reason="Use direct calls to the ReourceManager.")

geonode/services/forms.py

@@ -22,15 +22,15 @@
 from django import forms
 from django.core.exceptions import ValidationError
 from django.utils.translation import gettext_lazy as _
-import taggit
+from taggit.forms import TagField
 
 from geonode.security.auth_handlers import BasicAuthHandler
 from geonode.security.auth_registry import auth_handler_registry
 from geonode.security.models import AuthConfig
 
-from . import enumerations
-from .models import Service, get_service_type_choices
-from .serviceprocessors import get_service_handler
+from geonode.services import enumerations
+from geonode.services.models import Service, get_service_type_choices
+from geonode.services.serviceprocessors import get_service_handler
 from geonode.utils import is_safe_url
 
 logger = logging.getLogger(__name__)
@@ -72,10 +72,6 @@ def clean_url(self):
 
         if not is_safe_url(proposed_url):
             raise ValidationError(_("Invalid URL provided"))
-
-        existing = Service.objects.filter(base_url=proposed_url).exists()
-        if existing:
-            raise ValidationError(_("Service %(url)s is already registered"), params={"url": proposed_url})
         return proposed_url
 
     def clean(self):
@@ -129,7 +125,7 @@ class ServiceForm(forms.ModelForm):
     )
     description = forms.CharField(label=_("Description"), widget=forms.Textarea(attrs={"cols": 60}))
     abstract = forms.CharField(label=_("Abstract"), widget=forms.Textarea(attrs={"cols": 60}))
-    keywords = taggit.forms.TagField(required=False)
+    keywords = TagField(required=False)
 
     class Meta:
         model = Service

geonode/services/migrations/0061_alter_service_base_url.py

@@ -0,0 +1,19 @@
+# Generated by Django 5.2.13 on 2026-07-02 10:00
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("services", "0060_alter_service_type"),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name="service",
+            name="base_url",
+            field=models.URLField(db_index=True),
+        ),
+    ]
+

geonode/services/models.py

@@ -31,7 +31,7 @@
 from geonode.layers.enumerations import GXP_PTYPES
 from geonode.people.enumerations import ROLE_VALUES
 from geonode.services.serviceprocessors import get_available_service_types
-from . import enumerations
+from geonode.services import enumerations
 
 
 def get_service_type_choices():
@@ -56,8 +56,9 @@ class Service(ResourceBase):
             (enumerations.OPENGEOPORTAL, _("OpenGeoPortal")),
         ),
     )
-    # with service, version and request etc stripped off
-    base_url = models.URLField(unique=True, db_index=True)
+    # URL with service, version and request etc stripped off
+    # Intentionally non-unique: the same endpoint can be registered by different owners/use-cases.
+    base_url = models.URLField(db_index=True)
     version = models.CharField(max_length=100, null=True, blank=True)
     # Should force to slug?
     name = models.CharField(max_length=255, unique=True, db_index=True)

geonode/services/serviceprocessors/__init__.py

@@ -18,31 +18,43 @@
 #########################################################################
 import logging
 
-from django.conf import settings
 from geonode.services import enumerations
-from django.core.cache import caches
-from geonode.services.serviceprocessors.registry import service_type_registry
+from geonode.services.serviceprocessors.cache import service_handler_cache
+from geonode.services.serviceprocessors.registry import service_type_registry
 
-service_cache = caches["services"]
 logger = logging.getLogger(__name__)
 
 
-def get_available_service_types():
-    return service_type_registry.get_available_service_types()
+def get_service_cache_key(base_url, service_type=enumerations.AUTO, service_id=None, auth=None, auth_config=None):
+    return service_handler_cache.get_key(
+        base_url, service_type=service_type, service_id=service_id, auth=auth, auth_config=auth_config
+    )
+
+
+def get_available_service_types():
+    return service_type_registry.get_available_service_types()
 
 
 def get_service_handler(base_url, service_type=enumerations.AUTO, service_id=None, *args, **kwargs):
     """Return the appropriate remote service handler for the input URL.
     If the service type is not explicitly passed in it will be guessed from
     """
 
-    if entry := service_cache.get(base_url):
+    cache_key = get_service_cache_key(
+        base_url,
+        service_type=service_type,
+        service_id=service_id,
+        auth=kwargs.get("auth"),
+        auth_config=kwargs.get("auth_config"),
+    )
+
+    if entry := service_handler_cache.get(cache_key):
         return entry
 
-    handler = service_type_registry.get_handler_class(service_type)
+    handler = service_type_registry.get_handler_class(service_type)
     try:
         service_handler = handler(base_url, service_id, *args, **kwargs)
-        service_cache.set(service_handler.url, service_handler, settings.SERVICE_CACHE_EXPIRATION_TIME)
+        service_handler_cache.set(cache_key, service_handler)
     except Exception as e:
         logger.exception(e)
         logger.exception(msg=f"Could not parse service {base_url}")

geonode/services/serviceprocessors/arcgis.py

@@ -20,9 +20,11 @@
 import os
 import logging
 import traceback
-
+from collections import namedtuple
 from uuid import uuid4
 
+from arcrest import MapService as ArcMapService, ImageService as ArcImageService
+
 from django.conf import settings
 from django.db import transaction
 from django.utils.translation import gettext_lazy as _
@@ -31,16 +33,9 @@
 from geonode import GeoNodeException
 from geonode.base.bbox_utils import BBOXHelper
 from geonode.harvesting.models import Harvester
+from geonode.services import enumerations, models, utils
+from geonode.services.serviceprocessors import base
 
-from arcrest import MapService as ArcMapService, ImageService as ArcImageService
-
-from .. import enumerations
-from ..enumerations import INDEXED
-from .. import models
-from .. import utils
-from . import base
-
-from collections import namedtuple
 
 logger = logging.getLogger(__name__)
 
@@ -86,7 +81,7 @@ def __init__(self, url, geonode_service_id=None, *args, **kwargs):
         #     extent['ymax']
         # ])
 
-        self.indexing_method = INDEXED
+        self.indexing_method = enumerations.INDEXED
         self.name = slugify(self.url)[:255]
         self.title = str(_title).encode("utf-8", "ignore").decode("utf-8")
 
@@ -110,39 +105,44 @@ def create_geonode_service(self, owner, parent=None):
         :type owner: geonode.people.models.Profile
 
         """
-        with transaction.atomic():
-            instance = models.Service.objects.create(
-                uuid=str(uuid4()),
-                base_url=self.url,
-                type=self.service_type,
-                method=self.indexing_method,
-                owner=owner,
-                metadata_only=True,
-                version=str(self.parsed_service._json_struct.get("currentVersion", 0.0))
-                .encode("utf-8", "ignore")
-                .decode("utf-8"),
-                name=self.name,
-                title=self.title,
-                abstract=str(self.parsed_service._json_struct.get("serviceDescription"))
-                .encode("utf-8", "ignore")
-                .decode("utf-8")
-                or _("Not provided"),
-            )
-            service_harvester = Harvester.objects.create(
-                name=self.name,
-                default_owner=owner,
-                scheduling_enabled=False,
-                remote_url=instance.service_url,
-                delete_orphan_resources_automatically=True,
-                harvester_type=enumerations.HARVESTER_TYPES[self.service_type],
-                harvester_type_specific_configuration=self.get_harvester_configuration_options(),
-            )
-            if service_harvester.update_availability():
-                service_harvester.initiate_update_harvestable_resources()
-            else:
-                logger.exception(GeoNodeException("Could not reach remote endpoint."))
-            instance.harvester = service_harvester
 
+        def _create(unique_name):
+            with transaction.atomic():
+                new_instance = models.Service.objects.create(
+                    uuid=str(uuid4()),
+                    base_url=self.url,
+                    type=self.service_type,
+                    method=self.indexing_method,
+                    owner=owner,
+                    metadata_only=True,
+                    version=str(self.parsed_service._json_struct.get("currentVersion", 0.0))
+                    .encode("utf-8", "ignore")
+                    .decode("utf-8"),
+                    name=unique_name,
+                    title=self.title,
+                    abstract=str(self.parsed_service._json_struct.get("serviceDescription"))
+                    .encode("utf-8", "ignore")
+                    .decode("utf-8")
+                    or _("Not provided"),
+                )
+                new_harvester = Harvester.objects.create(
+                    name=unique_name,
+                    default_owner=owner,
+                    scheduling_enabled=False,
+                    remote_url=new_instance.service_url,
+                    delete_orphan_resources_automatically=True,
+                    harvester_type=enumerations.HARVESTER_TYPES[self.service_type],
+                    harvester_type_specific_configuration=self.get_harvester_configuration_options(),
+                )
+                if new_harvester.update_availability():
+                    new_harvester.initiate_update_harvestable_resources()
+                else:
+                    logger.exception(GeoNodeException("Could not reach remote endpoint."))
+                new_instance.harvester = new_harvester
+                return new_instance
+
+        instance = base.create_with_unique_name(self.name, _create)
+        self.name = instance.name
         self.geonode_service_id = instance.id
         return instance
 
@@ -235,7 +235,7 @@ def __init__(self, url, geonode_service_id=None, *args, **kwargs):
         #     extent['ymax']
         # ])
 
-        self.indexing_method = INDEXED
+        self.indexing_method = enumerations.INDEXED
         self.name = slugify(self.url)[:255]
         self.title = _title
 

geonode/services/serviceprocessors/base.py

@@ -22,14 +22,14 @@
 import logging
 
 from django.conf import settings
+from django.db import IntegrityError
 
 from geonode.utils import check_ogc_backend
 from geonode import GeoNodeException, geoserver
 from geonode.harvesting.tasks import harvest_resources
-from geonode.harvesting.models import AsynchronousHarvestingSession
+from geonode.harvesting.models import AsynchronousHarvestingSession, Harvester
 
-from .. import models
-from .. import enumerations
+from geonode.services import models, enumerations
 
 if check_ogc_backend(geoserver.BACKEND_PACKAGE):
     from geonode.geoserver.helpers import gs_catalog as catalog
@@ -51,6 +51,47 @@ def get_geoserver_cascading_workspace(create=True):
     return workspace
 
 
+def build_unique_resource_name(name, max_length=255):
+    """Return a Service/Harvester name that is unique in both models."""
+    max_length = max(1, int(max_length))
+    candidate = (name or "service")[:max_length]
+    base_name = candidate
+    idx = 1
+    while models.Service.objects.filter(name=candidate).exists() or Harvester.objects.filter(name=candidate).exists():
+        suffix = f"-{idx}"
+        if len(suffix) >= max_length:
+            # When max_length is tiny, keep the most specific part of the suffix.
+            candidate = suffix[-max_length:]
+        else:
+            prefix_len = max_length - len(suffix)
+            candidate = f"{base_name[:prefix_len]}{suffix}"
+        idx += 1
+    return candidate
+
+
+def create_with_unique_name(name, create_fn, max_length=255, max_attempts=3):
+    """Call create_fn(unique_name) with a freshly generated unique candidate name,
+    retrying under a new candidate if a concurrent registration raced us to the
+    same name and tripped the DB's uniqueness constraint on Service.name.
+
+    build_unique_resource_name's own existence check is not atomic with the
+    caller's actual create(), so two concurrent registrations can still land
+    on the same candidate; this retries the whole attempt with a fresh name
+    instead of surfacing the resulting IntegrityError to the caller.
+
+    :arg create_fn: callable invoked with the candidate name; must perform the
+        actual object creation(s) and raise IntegrityError if the name turned
+        out to already be taken.
+    """
+    for attempt in range(max_attempts):
+        candidate = build_unique_resource_name(name, max_length=max_length)
+        try:
+            return create_fn(candidate)
+        except IntegrityError:
+            if attempt == max_attempts - 1:
+                raise
+
+
 class ServiceHandlerBase(object):  # LGTM: @property will not work in old-style classes
     """Base class for remote service handlers