You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm converting pipelines from SDK 1.8.6 to 2.9.0 and having issues when trying to pass artifacts from components outside of dsl condition blocks to components within them. I created a simplified example to show the error I'm having.
Environment
How did you deploy Kubeflow Pipelines (KFP)? Deployed on a Kubernetes cluster using Kubeflow manifests.
The pipeline fails during execution with an error related to resolving input artifacts in conditional branches.
Expected result
The print_pokemon_message component should receive the choose_txt_artifact output from start_comp and print the appropriate message.
The pipeline fails with an error when trying to resolve the input artifact pipelinechannel--start-comp-choose_txt_artifact within the conditional branches:
failed to resolve inputs: failed to resolve input artifact pipelinechannel--start-comp-choose_txt_artifact with spec component_input_artifact:"pipelinechannel--start-comp-choose_txt_artifact": component input artifact not implemented yet
I believe passing artifacts into components inside conditional branches may not be working as expected.
Materials and Reference
parentDAGinputparametersmap[pipelinechannel--choose-pokemon-starter-Output:string_value:"Grass"]
KFPdriver:driver.DAG(pipelineName=pokemon-pipeline,runID=[REDACTED],task="condition-4",component="comp-condition-4",dagExecutionID=[REDACTED],componentSpec) failed:failedtoresolveinputs:failedtoresolveinputartifactpipelinechannel--start-comp-choose_txt_artifactwithspeccomponent_input_artifact:"pipelinechannel--start-comp-choose_txt_artifact":componentinputartifactnotimplementedyettime="2024-10-08T15:37:18.142Z"level=infomsg="sub-process exited"argo=trueerror="<nil>"time="2024-10-08T15:37:18.142Z"level=errormsg="cannot save parameter /tmp/outputs/execution-id"argo=trueerror="open /tmp/outputs/execution-id: no such file or directory"time="2024-10-08T15:37:18.142Z"level=errormsg="cannot save parameter /tmp/outputs/iteration-count"argo=trueerror="open /tmp/outputs/iteration-count: no such file or directory"time="2024-10-08T15:37:18.142Z"level=errormsg="cannot save parameter /tmp/outputs/condition"argo=trueerror="open /tmp/outputs/condition: no such file or directory"
Impacted by this bug? Give it a 👍.
The text was updated successfully, but these errors were encountered:
I'm converting pipelines from SDK 1.8.6 to 2.9.0 and having issues when trying to pass artifacts from components outside of dsl condition blocks to components within them. I created a simplified example to show the error I'm having.
Environment
Steps to reproduce
Use the following code to create a pipeline:
The pipeline fails during execution with an error related to resolving input artifacts in conditional branches.
Expected result
The print_pokemon_message component should receive the choose_txt_artifact output from start_comp and print the appropriate message.
The pipeline fails with an error when trying to resolve the input artifact pipelinechannel--start-comp-choose_txt_artifact within the conditional branches:
I believe passing artifacts into components inside conditional branches may not be working as expected.
Materials and Reference
Impacted by this bug? Give it a 👍.
The text was updated successfully, but these errors were encountered: