We’re currently setting up our CI/CD to automatize the deployment of AI Bridge into NGC.
We’re using the latest NGC CLI (3.44.0).
We’re almost finished, however, we’re having an error when calling: “ngc registry collection update”, however, the very same command with create works as expected.
Bug/Problem description:
1 – At the end of our workflow/pipeline we verify if a collection exists on NGC, if it doesn’t (first time we ran the workflow) we call: ‘ngc registry collection create’
Command output:
Command: ngc registry collection --debug create --org xxx-xxx --team aaa --display-name XXX --short-desc XXX --overview-filename /tmp/overview.md --logo https://download/collection.png --category INTELLIGENT_VIDEO_ANALYTICS --label AI --label Inference --label Infrastructure --label Multi-Node --label Multi-Arch --label X86 --label ARM --add-chart xxx-xxx/aaa/bbb --add-resource xxx-xxxx/aaa/bbb --add-resource xxx-xxx/aaa/ccc --add-resource xxx-xxx/aaa/ddd --add-image xxx-xxx/aaa/eee --add-image xxx-xxx/aaa/fff --add-image xxx-xxx/aaa/ggg --add-image xxx-xxx/aaa/hhh --add-image xxx-xxx/aaa/iii --add-image xxx-xxx/aaa/jjj --add-image xxx-xxx/aaa/aibridge-kkk --add-image xxx-xxx/aaa/lll --add-image xxx-xxx/aaa/mmm --add-image xxx-xxx/aaa/nnn --publisher XXXX --built-by XXXXX xxx-xxx/aaa/ooo
os: Linux-5.15.0-112-generic
ngc-cli version: NGC CLI 3.44.0
Run “ngc diag all” to view more detailed information.
And works as expected:
However, if we retry our workflow/pipeline and we call: ngc registry collection update
Command output:
Command: ngc registry collection --debug update --org xxx-xxx --team aaa --display-name XXX --short-desc XXX --overview-filename /tmp/overview.md --logo https://download/collection.png --category INTELLIGENT_VIDEO_ANALYTICS --label AI --label Inference --label Infrastructure --label Multi-Node --label Multi-Arch --label X86 --label ARM --add-chart xxx-xxx/aaa/bbb --add-resource xxx-xxxx/aaa/bbb --add-resource xxx-xxx/aaa/ccc --add-resource xxx-xxx/aaa/ddd --add-image xxx-xxx/aaa/eee --add-image xxx-xxx/aaa/fff --add-image xxx-xxx/aaa/ggg --add-image xxx-xxx/aaa/hhh --add-image xxx-xxx/aaa/iii --add-image xxx-xxx/aaa/jjj --add-image xxx-xxx/aaa/aibridge-kkk --add-image xxx-xxx/aaa/lll --add-image xxx-xxx/aaa/mmm --add-image xxx-xxx/aaa/nnn --publisher XXXX --built-by XXXXX xxx-xxx/aaa/ooo
os: Linux-5.15.0-112-generic
ngc-cli version: NGC CLI 3.44.0
Run “ngc diag all” to view more detailed information.
Failed to export traces to prod.otel.kaizen.nvidia.com:8282, error code: StatusCode.UNKNOWN
Traceback (most recent call last):
File “opentelemetry/exporter/otlp/proto/grpc/exporter.py”, line 262, in _export
self._client.Export(
File “grpc/_channel.py”, line 1176, in call
File “grpc/_channel.py”, line 1005, in _end_unary_response_blocking
grpc._channel._InactiveRpcError: <_InactiveRpcError of RPC that terminated with:
status = StatusCode.UNKNOWN
details = “Received http2 header with status: 413”
debug_error_string = “UNKNOWN:Error received from peer {created_time:“2024-06-19T15:20:59.616801239+00:00”, grpc_status:2, grpc_message:“Received http2 header with status: 413”}”
We do get some ‘python’ errors from the server-side.
This was not happening in the past (we fear is a regression on your side),…