skypilot-org / skypilot

SkyPilot: Run AI and batch jobs on any infra (Kubernetes or 12+ clouds). Get unified execution, cost savings, and high GPU availability via a simple interface.
https://skypilot.readthedocs.io
Apache License 2.0
6.56k stars 469 forks source link

[UX/AWS] When SSO expired, `sky status` shows a lot SSO error #3967

Open Michaelvll opened 2 days ago

Michaelvll commented 2 days ago

When AWS SSO session expired, sky status can show a lot of SSO error if the catalog needs to be updated (reached the update frequency).

  File "/opt/conda/envs/sky-dev/lib/python3.10/site-packages/botocore/tokens.py", line 95, in _refresh
    self._protected_refresh()
  File "/opt/conda/envs/sky-dev/lib/python3.10/site-packages/botocore/tokens.py", line 122, in _protected_refresh
    raise TokenRetrievalError(
botocore.exceptions.TokenRetrievalError: Error when retrieving token from sso: Token has expired and refresh failed
D 09-20 08:16:38 aws_catalog.py:146] Failed to get AWS user identity. Using the latest mapping file for user 'b710cff4'.

Version & Commit info:

Michaelvll commented 2 days ago

For record from #3754

A user reported that, even with sky status, it requires AWS SSO login once the authentication is expired. This is unexpected, as the sky status should be considered to be not needing to connect to the clouds.