Python notebook error: Failed to start the kernel

[zhaohuifan]

Hi, 

I just started to build my first ML project using Dataiku. I got a kernel error whenever I try to open and edit a python code notebook:

Traceback (most recent call last):
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/python36.packages/tornado/web.py", line 1592, in _execute
result = yield result
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/python36.packages/tornado/gen.py", line 1133, in run
value = future.result()
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/python36.packages/tornado/gen.py", line 326, in wrapper
yielded = next(result)
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/dku-jupyter/packages/notebook/services/sessions/handlers.py", line 67, in post
session_virtual_path = sm.get_session_virtual_path(path=path, user=self.get_current_user())
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/dku-jupyter/packages/notebook/dataiku/sessionmanager.py", line 24, in get_session_virtual_path
return self.get_session_virtual_path_(head, tail, user)
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/dku-jupyter/packages/notebook/dataiku/sessionmanager.py", line 12, in get_session_virtual_path_
"user" : user
File "/mnt/dss/dataiku/dss_installdir/dataiku-dss-10.0.7/dku-jupyter/packages/notebook/dataiku/__init__.py", line 33, in pintercom_json_post
raise Exception("Failed privileged call (%s): %s" % (path, _get_error_message(err_data).encode("utf8")))
Exception: Failed privileged call (jupyter/get-session-virtual-path): b'No user impersonation rule matched for user <AC:user:xxx.xxx@xxx.com>'

What does the "user impersonation rule" mean? How to set it up? 

Thank you.

---

Operating system used: Windows 10

[RoyE · Accepted solution]

Hello!

For more information on User Isolation Framework please refer to our documentation described here.

This error is occurring due to a setup issue with your UIF and your user is unable to be mapped properly therefore you do not have the correct privileges to run the task. 

You will need to reach out to an administrator of your instance to double check the UIF identity mapping as your user is not mapped properly. 

Please let me know if you have any further questions!