CLI shows a puppet upload is successful even though server side has an exception
The CLI incorrectly reports that an upload is successful even though the server may have had an exception that causes the upload to not be imported into the repo correctly.
This definitly affects the puppet module, but may affect all plugins. To reproduce:
0) create a puppet repo
1) find a valid puppet module
2) edit the tar.gz file and remove all metadata.json files (this invalidates the puppet module)
3) upload it to pulp
4) observe that it will show no errors
5) observe that there will be a server side traceback saying the metadata.json could not be found or is invalid
I expected that since the import of the uploaded module into the repo did not work that the CLI would reflect that.
The issue is that the CLI does not poll based on the async task ID given returned from the call to /pulp/api/v2/repositories/<repoid>/actions/import_upload/ It should poll and only show the user that the upload was successful if that task is successful.
Updated by email@example.com over 7 years ago
seems like this kind of bug already was reported and was closed with this comment https://bugzilla.redhat.com/show_bug.cgi?id=1104157#c1
Is it still the case?
Updated by bmbouter over 7 years ago
I feel this is a valid bug because the last step of uploading is an import into a repo and if the import fails then the upload also needs to fail. The succeeded message indicates to the user that the uploaded content is in the repo when in fact, it is not.