[Bug] In the continuous integration the notebook test frequently fails

[jan-janssen]

Maybe it helps to use a flux based kernel, rather than starting flux first and then calling papermill. A template for a flux kernel is available in https://github.com/pyiron-dev/pyiron-script-job-queue/blob/main/binder/kernel.json .

[jan-janssen]

The surprising part is that the notebook integration tests work fine even though they also use flux and are started in the same way.

[jan-janssen]

I tried both - reordering the execution of the jupyter notebooks #725 as well as creating a custom kernel #724 but neither resolved the issue.

[jan-janssen]

Maybe this bug was resolved by #738

[jan-janssen]

While the issue was not addressed explicitly, it seems to no longer show up so frequently, so let's hope it was just and issue on the continuous integration side of Github.