Error when starting JupyterHub instance after long period of inactivity
Summary
Starting a JupyterHub server after a long period of inactivity (48 hrs) resulted in a 500 error from the JupyterHub server
Further review of pod logs show the following error:
Exec lifecycle hook ([git clone https://gitlab.com/gitlab-org/nurtch-demo.git DevOps-Runbook-Demo]) for Container "notebook" in Pod "jupyter-danielgruesso_gitlab-managed-apps(f7747ddc-db8f-11e8-a391-42010a8000d1)" failed - error: command 'git clone https://gitlab.com/gitlab-org/nurtch-demo.git DevOps-Runbook-Demo' exited with 128: fatal: destination path 'DevOps-Runbook-Demo' already exists and is not an empty directory. , message: "fatal: destination path 'DevOps-Runbook-Demo' already exists and is not an empty directory.\n": PostStartHookError
Steps to reproduce
- Deploy JupyterHub to a k8s cluster via the GitLab k8s integration
- Login to JupyterHub to verify it's working and you can use
- Don't use the server for 24 hrs
- Attempting to use after 24 hrs will reveal the error
Example Project
http://jupyter.35.238.195.231.nip.io/
from