Hi All,
I am asking for your help to resolve the “503:Service Unavailable. Your server appears to be down” error.
We installed the latest version of Jupyterhub to test it on a small cluster composed of only two nodes: server (107) and computing node (106). The Jupyterhub ran very well with all users with no SSL certificates within a virtual Conda environment. However, when we wanted to test the PBS BatchSpawner (GitHub - jupyterhub/batchspawner: Custom Spawner for Jupyterhub to start servers in batch scheduled systems); it resulted in the 503 error. The Conda environment with the BatchSpawner has been installed on a shared network folder. We checked all answers posted online to resolve this issue but with no luck so far!
The Environment versions:
IPython : 7.31.1
ipykernel : 6.4.1
ipywidgets : not installed
jupyter_client : 7.1.2
jupyter_core : 4.9.1
jupyter_server : 1.13.5
jupyterlab : 3.2.1
nbclient : 0.5.3
nbconvert : 6.3.0
nbformat : 5.1.3
notebook : 6.4.6
qtconsole : not installed
traitlets : 5.1.1
configurable-http-proxy: 4.5.1
PBS: 19.1.3
The config file of Jupyterhub:
c.Application.log_level = ‘DEBUG’
import batchspawner
c.JupyterHub.spawner_class = ‘batchspawner.PBSSpawner’
c.Spawner.http_timeout = 180
c.BatchSpawnerBase.req_nprocs = ‘2’
c.BatchSpawnerBase.req_queue = ‘hpcc’
c.BatchSpawnerBase.req_host = ‘kcn107’
c.BatchSpawnerBase.req_runtime = ‘00:30:00’
c.BatchSpawnerBase.req_memory = ‘10gb’
c.PBSSpawner.batch_script = ‘’’#!/bin/bash
#PBS -q {queue}@{host}
#PBS -l walltime={runtime}
#PBS -l select=1:ncpus={nprocs}:mem={memory}
#PBS -N jupyterhub-singleuser
#PBS -v {keepvars}
#PBS -o /home/{username}/test.o
#PBS -e /home/{username}/test.e
cd $PBS_0_WORKDIR
echo $HOSTNAME
env
{cmd}
‘’’
c.PBSSpawner.state_exechost_exp = ‘kcn106’
c.JupyterHub.bind_url = ‘http://:8000’
c.JupyterHub.hub_connect_ip = ‘10.113.41.107’
c.JupyterHub.hub_ip = ‘kcn107’
c.BatchSpawnerBase.ip = ‘0.0.0.0’
The JupyterHub log file is attached. No changes have been made to the BatchSpawner config file. Your help is much appreciated - Thank you,