@ancient-wolf-19325 How can I prevent Flyte from retrying tasks on the same node

Flyte does not provide a built-in mechanism to prevent task retries from occurring on the same Kubernetes node; task placement is managed by Kubernetes, not Flyte. You can influence node selection by specifying node affinities or tolerations in the task's pod spec, but Flyte itself does not guarantee that retries will be scheduled on a different node than previous attempts. This limitation is not explicitly addressed in the official documentation or API references, and no direct configuration for "retry on a different node" exists in the provided data sources. For more control, you would need to use Kubernetes features such as pod anti-affinity in your task configuration, but this is outside Flyte's native retry logic scope. Would you like more detail on how to use Kubernetes pod affinity/anti-affinity with Flyte tasks? Sources: - https://github.com/flyteorg/flyte/issues/5217 - https://github.com/flyteorg/flyte/blob/master/docs/user_guide/flyte_fundamentals/optimizing_tasks.md 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.