X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/7334ee9ee6350f2b5f0384d8166e7c4f58d86864..540b72d62a94015f116ba077e279a5f10d666778:/sdk/cwl/arvados_cwl/arvdocker.py

diff --git a/sdk/cwl/arvados_cwl/arvdocker.py b/sdk/cwl/arvados_cwl/arvdocker.py
index 04e2a4cffc..f5e67a6649 100644
--- a/sdk/cwl/arvados_cwl/arvdocker.py
+++ b/sdk/cwl/arvados_cwl/arvdocker.py
@@ -17,9 +17,6 @@ import arvados.commands.keepdocker
 
 logger = logging.getLogger('arvados.cwl-runner')
 
-cached_lookups = {}
-cached_lookups_lock = threading.Lock()
-
 def determine_image_id(dockerImageId):
     for line in (
         subprocess.check_output(  # nosec
@@ -56,10 +53,16 @@ def determine_image_id(dockerImageId):
     return None
 
 
-def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid,
-                         force_pull, tmp_outdir_prefix, match_local_docker):
+def arv_docker_get_image(api_client, dockerRequirement, pull_image, runtimeContext):
     """Check if a Docker image is available in Keep, if not, upload it using arv-keepdocker."""
 
+    project_uuid = runtimeContext.project_uuid
+    force_pull = runtimeContext.force_docker_pull
+    tmp_outdir_prefix = runtimeContext.tmp_outdir_prefix
+    match_local_docker = runtimeContext.match_local_docker
+    copy_deps = runtimeContext.copy_deps
+    cached_lookups = runtimeContext.cached_docker_lookups
+
     if "http://arvados.org/cwl#dockerCollectionPDH" in dockerRequirement:
         return dockerRequirement["http://arvados.org/cwl#dockerCollectionPDH"]
 
@@ -69,22 +72,29 @@ def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid
         if hasattr(dockerRequirement, 'lc'):
             dockerRequirement.lc.data["dockerImageId"] = dockerRequirement.lc.data["dockerPull"]
 
-    global cached_lookups
-    global cached_lookups_lock
-    with cached_lookups_lock:
-        if dockerRequirement["dockerImageId"] in cached_lookups:
-            return cached_lookups[dockerRequirement["dockerImageId"]]
+    if dockerRequirement["dockerImageId"] in cached_lookups:
+        return cached_lookups[dockerRequirement["dockerImageId"]]
 
     with SourceLine(dockerRequirement, "dockerImageId", WorkflowException, logger.isEnabledFor(logging.DEBUG)):
         sp = dockerRequirement["dockerImageId"].split(":")
         image_name = sp[0]
         image_tag = sp[1] if len(sp) > 1 else "latest"
 
-        images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
+        out_of_project_images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
                                                                 image_name=image_name,
-                                                                image_tag=image_tag)
+                                                                image_tag=image_tag,
+                                                                project_uuid=None)
 
-        if images and match_local_docker:
+        if copy_deps:
+            # Only images that are available in the destination project
+            images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
+                                                                    image_name=image_name,
+                                                                    image_tag=image_tag,
+                                                                    project_uuid=project_uuid)
+        else:
+            images = out_of_project_images
+
+        if match_local_docker:
             local_image_id = determine_image_id(dockerRequirement["dockerImageId"])
             if local_image_id:
                 # find it in the list
@@ -98,15 +108,26 @@ def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid
                     # force re-upload.
                     images = []
 
+                for i in out_of_project_images:
+                    if i[1]["dockerhash"] == local_image_id:
+                        found = True
+                        out_of_project_images = [i]
+                        break
+                if not found:
+                    # force re-upload.
+                    out_of_project_images = []
+
         if not images:
-            # Fetch Docker image if necessary.
-            try:
-                result = cwltool.docker.DockerCommandLineJob.get_image(dockerRequirement, pull_image,
-                                                              force_pull, tmp_outdir_prefix)
-                if not result:
-                    raise WorkflowException("Docker image '%s' not available" % dockerRequirement["dockerImageId"])
-            except OSError as e:
-                raise WorkflowException("While trying to get Docker image '%s', failed to execute 'docker': %s" % (dockerRequirement["dockerImageId"], e))
+            if not out_of_project_images:
+                # Fetch Docker image if necessary.
+                try:
+                    dockerjob = cwltool.docker.DockerCommandLineJob(None, None, None, None, None, None)
+                    result = dockerjob.get_image(dockerRequirement, pull_image,
+                                                                  force_pull, tmp_outdir_prefix)
+                    if not result:
+                        raise WorkflowException("Docker image '%s' not available" % dockerRequirement["dockerImageId"])
+                except OSError as e:
+                    raise WorkflowException("While trying to get Docker image '%s', failed to execute 'docker': %s" % (dockerRequirement["dockerImageId"], e))
 
             # Upload image to Arvados
             args = []
@@ -125,20 +146,14 @@ def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid
 
             images = arvados.commands.keepdocker.list_images_in_arv(api_client, 3,
                                                                     image_name=image_name,
-                                                                    image_tag=image_tag)
+                                                                    image_tag=image_tag,
+                                                                    project_uuid=project_uuid)
 
         if not images:
             raise WorkflowException("Could not find Docker image %s:%s" % (image_name, image_tag))
 
         pdh = api_client.collections().get(uuid=images[0][0]).execute()["portable_data_hash"]
 
-        with cached_lookups_lock:
-            cached_lookups[dockerRequirement["dockerImageId"]] = pdh
+        cached_lookups[dockerRequirement["dockerImageId"]] = pdh
 
     return pdh
-
-def arv_docker_clear_cache():
-    global cached_lookups
-    global cached_lookups_lock
-    with cached_lookups_lock:
-        cached_lookups = {}