The error "failed to create workflow in propeller etcdserver: request is too large" is typically caused by exceeding the size limit for requests to etcd, which is used by Flyte for storing workflow metadata. This can happen if the workflow or its components are too large, such as having a large number of nodes or large data payloads. To address this, you can enable offloading of static workflow information from the Custom Resource Definition (CRD) to blob storage by setting the

useOffloadedWorkflowClosure

parameter to

true

. This reduces the size of the data stored in etcd and can help avoid the error (source). Would you like more detail on how to configure this setting? Sources: - https://discuss.flyte.org/t/16659081/hi-just-tried-updating-from-1-9-to-1-11-and-my-workflows-are/ 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

hey @thousands-car-79657 was the above helpful for you?

Hi David! It seems helpful, but I haven’t tried the suggested solution. Will do and report back

To configure

useOffloadedWorkflowClosure

in the Flyte Helm chart, you need to set this parameter in the FlyteAdmin configuration. This involves editing the configuration file to enable offloading of static workflow information to a blob store. Here's a general approach: 1. Locate the FlyteAdmin Configuration: This is typically found in the Helm chart values or a configuration file used by FlyteAdmin. 2. Set the Parameter: Add or update the

useOffloadedWorkflowClosure

parameter to

true

in the configuration. 3. Deploy the Changes: After updating the configuration, redeploy the Flyte Helm chart to apply the changes. For more detailed steps, you might need to refer to the specific Helm chart documentation or configuration files used in your deployment. If you need further assistance, please let me know! Sources: - https://docs.flyte.org/en/latest/deployment/configuration/performance.html#offloading-static-workflow-information-from-crd 👍 0 button 👎 1 button 💡 Hint: Mention @ancient-wolf-19325 in the thread for followups.

@ancient-wolf-19325 sorry, I meant specifically how to configure

useOffloadedWorkflowClosure

in the

configmap

?

To configure

useOffloadedWorkflowClosure

in the

configmap

, you need to update the FlyteAdmin configuration within the Kubernetes ConfigMap. Here's how you can do it: 1. Edit the ConfigMap: Locate the ConfigMap used by FlyteAdmin. This is typically named something like

flyte-admin-config

or similar, depending on your setup. 2. Add the Configuration: Within the ConfigMap, add or update the

useOffloadedWorkflowClosure

setting under the appropriate section. It should look something like this:

apiVersion: v1
   kind: ConfigMap
   metadata:
     name: flyte-admin-config
     namespace: flyte
   data:
     flyteadmin_config.yaml: |
       adminServer:
         flyteadmin:
           useOffloadedWorkflowClosure: true

3. Apply the Changes: After editing the ConfigMap, apply the changes using

kubectl apply -f <configmap-file>.yaml

. 4. Restart FlyteAdmin: To ensure the changes take effect, restart the FlyteAdmin deployment. You can do this with:

kubectl rollout restart deployment <flyteadmin-deployment-name> -n flyte

This configuration will enable the offloading of static workflow information to a blob store, reducing the size of the data stored in etcd. Would you like more detailed guidance on any of these steps? Sources: - https://github.com/flyteorg/flyte/pull/3087 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.