Hi @kzimmerman!

In the case of session_length:

It was a problem of permissions, indeed. Thanks for spotting it. It's fixed now. Please, check that you can access now both mforns.session_length and mforns.session_length_2. When you confirm, I'll close this task.

Regarding other data sets:

All data sets that we want to make visible to other analytics-related employees (i.e. via Superset) should belong to the group analytics-privatedata-users.
And also, the file permission mode should be 750, so that no-one outside analytics-privatedata-users can read the data.

hdfs dfs -chgrp -R analytics-privatedata-users /path/to/dataset
hdfs dfs -chmod -R 750 /path/to/dataset

Please, could you pass this requirements to data set creators?

Thanks a lot!

@mforns I'm still getting the same error when I try to load mforns.session_length and mforns.session_length_2, unfortunately :(

I'll make sure my team is aware of those requirements for data set creators! Do you know if this is documented for reference?

@kzimmerman I've checked and I could not find your username (kzeta right?) in the analytics-privatedata-users group.
That's probably why you can not access the session length data.
We should add you there. Created a task: T272982

@kzimmerman https://wikitech.wikimedia.org/wiki/Analytics/Data_access#Access_Groups

We are trying to simplify the procedure, but essentially for Superset a user needs to:

• Login via Wikidev account (user/pass), with the username in the wmf or nda LDAP groups. This grants access to all the dashboards based on Druid since they are not authenticated.
• Membership of the analytics-privatedata-users group (even without ssh access, this is new) to use Presto, since the user in this case is proxied and authenticated up to Hive/HDFS.

Comments are welcome, let me know if there are parts of the docs that can be improved!

Edit: I misread the comments and I saw that you were referring to the earlier comment about data ownership etc.., apologies, but any feedback on the above doc is good as well :)

Thanks @mforns !