hey Flyte folks so i have been seen this weird breaking beha

Question

hey Flyte folks so i have been seen this weird breaking behavior in relaunching a workflow that has optional input FlyteFile ```Workflow myworkflow fastq2bam workflow failed RuntimeExecutionError max number of system retry attempts 11 10 exhausted Last known status message Failed to check Catalog for previous results unexpected artifactData trimmed fastqs type map value type lt union type lt variants lt blob lt gt gt variants lt union type lt variants lt blob lt gt structure lt tag FlyteFilePath gt gt gt gt gt gt does not match any task output type map value type lt union type lt variants lt blob lt gt structure lt tag FlyteFilePath gt gt variants lt simple NONE structure lt tag none gt gt gt gt ``` unfortunately i am unable to replicate it in flyte sandbox yet Essentially when we relaunch a workflow due to failure in that workflow the successful task that has optional input is stuck in `Queued` state and flytepropeller is unable to resolve it did anyone encountered this issue

Jay Ganbat · Answer

maybe it is an issue on `datacatalog` side I also noticed that for individual relaunch optional types input is also getting ignored input becomes None and thus do not utilize caching

Kevin Su · Answer

how did you run the workflow at the beginning pyflyte run

Jay Ganbat · Answer

from flyteconsole

Kevin Su · Answer

so launch works but failed to relaunch

Jay Ganbat · Answer

yeah brand new launch will work but relaunch fails

Kevin Su · Answer

are you able to share the workflow code to reproduce we ll investigate it tomorrow

Jay Ganbat · Answer

so far no luck yet sweat smile yeah once i have something definitely will post it here

Dan Rammer (hamersaw) · Answer

This sounds very similar to <https github com flyteorg flyte issues 3901|this issue>