Open devinbost opened 2 hours ago
The issue you're experiencing seems to be related to the environment variable LANGFLOW_PORT
being set incorrectly, which is causing the application to interpret a URL as the port value. The LangFlow application expects the LANGFLOW_PORT
environment variable to be an integer, as it uses typer.Option
to parse this value [1].
Here are a few steps you can take to troubleshoot and resolve this issue:
Check Environment Variables: Ensure that the LANGFLOW_PORT
environment variable is set correctly in your Helm chart or Kubernetes configuration. It should be an integer, such as 7860
, and not a URL.
Helm Chart Configuration: Review your Helm chart's values.yaml
file and any other configuration files to ensure that the port is specified correctly. Look for any misconfigurations that might be setting the port to a URL.
Pod Environment: Verify the environment variables within the running pod. You can do this by describing the pod and checking the environment variables section to ensure LANGFLOW_PORT
is set to a valid integer.
Override Environment Variable: If necessary, you can manually set the LANGFLOW_PORT
environment variable in your deployment configuration to ensure it is set correctly.
Logs and Debugging: Check the logs for any additional errors or warnings that might provide more context on why the URL is being used as the port.
By ensuring that the LANGFLOW_PORT
is set correctly and is an integer, you should be able to resolve the issue with the port being assigned a URL.
To continue talking to Dosu, mention @dosu.
Bug Description
Deploying the helm chart with the latest code (main branch) results in a strange issue where the langflow service URL is substituted for the port value. There were no changes in the helm code, which leads me to believe that a change in the LangFlow source is responsible.
As proof of this, the URL matches the langflow service (top line):
I'm not finding any clues on the pod itself.... It appears that there's a missing environment variable, and as a consequence, the code is resolving the service endpoint. However, I can't figure out where that might be happening since all of the implementations of SettingService (and it's related modules) appear to be doing nothing creative with the port value.
I've tried disabling postgres (using sqlite instead), which had no effect, and I tried manually setting a PORT environment variable on the pod, which had no effect. So, I'm not sure how to even override this value.
Reproduction
replicated cluster shell (id from UI)
from terminalopenshift-postgres2
langflow-helm-charts/charts/langflow-ide
helm install langflow-ide . -f values.yaml --namespace langflow
Expected behavior
Pod should start normally.
Who can help?
@italojohnny @ogabrielluiz
Operating System
v4.14.0-okd
Langflow Version
master (latest)
Python Version
None
Screenshot
No response
Flow File
No response