on 07-30-2020 12:30 PM
Did anyone get the Model Serving Operator respectively the TensorFlow Serving Pipeline to work in SAP Data Intelligence?
My model gets loaded successfully:
[2020-07-30 10:29:09,000] Starting model deployment for artifacts ['6e0bb648-f88b-4e13-b2e8-e5362a566a5d'][2020-07-30 10:29:09,000] Preparing Deployment Spec[2020-07-30 10:29:09,000] Starting deployment for artifact 6e0bb648-f88b-4e13-b2e8-e5362a566a5d[2020-07-30 10:29:12,000] Response from api: {'id': 'ms-493d222a-1738-4d52-86fe-49ce595af073', 'specs': {'models': [{'namespace': 'default', 'modelName': '6e0bb648-f88b-4e13-b2e8-e5362a566a5d', 'modelVersion': '1', 'uri': 'https://storagegateway:14000/webhdfs/v1/shared/ml/artifacts/executions/1364090690294c3d953074a2a7ea8197/artifactproducer1_0.zip', 'labels': []}], 'modelRuntimeId': 'tf-1.14', 'resourcePlanId': 'basic', 'replicas': 1, 'ttl': '0m', 'enableHttpEndpoint': True, 'batchConfig': None, 'labels': [{'name': 'scenarioId-part1', 'value': 'XX-194d595c-ed7c-4857-8472-73a2efa049bb-XX'}, {'name': 'pipelineId-part1', 'value': 'XX-com.sap.dsp.templates.tensorflow_serving_template-XX'}, {'name': 'executionId-part1', 'value': 'XX-8e8f7dfde89646149e63a0e226b2bc76-XX'}], 'createdAt': '2020-07-30T10:29:27.727Z', 'updatedAt': '2020-07-30T10:29:27.727Z'}, 'status': {'state': 'PENDING', 'message': 'Initializing the deployment', 'updatedAt': '2020-07-30T10:29:28.312Z', 'expiredAt': None, 'replicaStatus': None}, 'endpoints': []}
But then there seems to be an internal error with the deployment:
[2020-07-30 10:29:12,000] Creating model server resources[2020-07-30 10:29:27,000] Finalizing the deployment resources creation[2020-07-30 10:29:34,000] Internal error. Please try again after some time or contact administrator.[2020-07-30 10:29:34,000] Shutting down Model Server ms-493d222a-1738-4d52-86fe-49ce595af073
This happens reproducably.
Many thank in advance
User | Count |
---|---|
84 | |
10 | |
10 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.