Closed jiyer2016 closed 6 years ago
You can ignore the No module named 'dask'
its coming from one of the dependencies s3fs
and it doesn't cause any issues.
The other one is different and I have seen it when using boto independently and its coming from boto. Maybe check the versions you are using?
Thanks, Would you be able to recommend the correct combinations of s3contents, s3fs and boto3 and botocore versions that are known to work well with each other ??
Here is the combination of libraries with which the issue happens. Perhaps - someone with a working setup can provide information on how different their setup as compared to the below:
boto3 1.9.23 botocore 1.12.23 ipykernel 4.10.0 ipython 7.0.1 ipython-genutils 0.2.0 ipython-sql 0.3.9 jupyter 1.0.0 jupyter-client 5.2.3 jupyter-console 6.0.0 jupyter-core 4.4.0 jupyterhub 0.9.4 jupyterlab 0.35.1 jupyterlab-server 0.2.0 notebook 5.7.0 s3contents 0.1.11 s3fs 0.1.5 s3transfer 0.1.13 tornado 5.1.1 traitlets 4.3.2
Can anybody help with this please ??
It's weird if it's a dependency since some of those are ==
. You can take a look at one of the CI builds to see whats passing for those: https://travis-ci.org/danielfrg/s3contents/jobs/441781489
Try adding this other settings to the ContentsManager: signature_version="s3v4"
.
Thanks for the response. The versions of the libraries are pretty much in sync with the travis logs. My setup uses minio s3 storage that supports the version 2 signature (and not v4). I added the following configuration - but get the same error - Thats strange.
c.S3ContentsManager.signature_version = "s3"
Do you see any problems working against a minio storage that supports only the v2 signature version ?
That could be the reason. The tests on Travis uses minio and uses s3v4
so imagine libraries only support new versions of the signature. I would recommend upgrading your minio.
Get the following error when starting up jupyter notebook:
Here is how my configuration looks:
Why would this happen ?