System error 13 when assigning product recognizer index to endpoint

Google AI Studio
,
1

I am trying to deploy product recognition index per this instruction in the guide:

And I am using this command to assign an index to the endpoint:
curl -sS -X POST -H “Authorization: Bearer $(gcloud auth print-access-token)” -H “Content-Type: application/json” https://visionai.googleapis.com/v1alpha1/projects/second-grail-471116-v8/locations/us-central1/retailEndpoints/516dbb61-9ea7-49f2-b769-9df91bdf24a5-sav-endpoint:deployRetailProductRecognitionIndex
-d ‘{
“retail_product_recognition_index”: “projects/second-grail-471116-v8/locations/us-central1/retailCatalogs/‘516dbb61-9ea7-49f2-b769-9df91bdf24a5’/retailProductRecognitionIndexes/516dbb61-9ea7-49f2-b769-9df91bdf24a5-indexv1”
}’

Prior to running that command, I confirm both the endpoint and index exist, which they do. Running the command starts an operation, but it always fails with this internal error response:

{
“name”: “projects/second-grail-471116-v8/locations/us-central1/operations/operation-1757045871056-63e061eff9581-41c538a8-31a76eef”,
“done”: true,
“error”: {
“code”: 13,
“message”: “an internal error has occurred”
}
}

Google Cloud support told me to reach out here for help. Thank you in advance.

3

Hello! Have you ever resolved this? I am stuck in the same command.

5

Hi Daniel! I have the same issue. Have you found a solution?

6

Hi, same here, it’s been almost a year, how has Google not resolved this bug?

7

It seems Google’s frontend accepts a valid request, but the running Vision AI server binary cannot resolve one of its own pinned internal model-package versions. You can actually see this error in the logs for the service:

SingletonNotReadyException
Unable to find an entry for build_time MPM version id
'1-7dd97eed_e9da37e2_c5acf36b_b89b2a23_40e37a7c'

I’ve filed a GCP support ticket about this.