New RunLLM response generated: >>>*# Question* I ...
# runllm-flyte-trial
a
New RunLLM response generated: >>>*# Question* I am using Flyte v1.13.0. If I run a workflow and due to some reason a task sends a large output (greater than maxDownloadMBs, I can see an error in the propeller logs:
Copy code
"error","msg":"Error when trying to reconcile workflow. Error [failed at Node[n0]. RuntimeExecutionError: failed during plugin execution, caused by: output file @[<>/n0/data/0/outputs.pb] is too large [19664776] bytes, max allowed [10485760] bytes]. Error Type[*errors.NodeErrorWithCause]"
But the workflow is stuck on running on flyte console….. I can see this issue to be similar but it doesn’t seem that this has been resolved. https://github.com/flyteorg/flyte/issues/381 # Answer The issue you're experiencing is indeed similar to the one described in Flyte Issue #381, where large outputs cause the workflow to get stuck instead of failing instantly. This issue has been acknowledged but not fully resolved. To address this, you can try increasing the
max-output-size-bytes
configuration in the FlytePropeller config. According to this discussion, you can modify the
flyte-propeller-config
ConfigMap to increase the
max-output-size-bytes
value and then restart the FlytePropeller pod. Would you like more detailed steps on how to modify the configuration? Sources: - Flyte Issue #381 - Flyte Discussion on Increasing Output Size Slack Link https://flyte-org.slack.com/archives/CP2HDHKE1/p1724409985.371659 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.