A new Superset release is available: https://github.com/apache/incubator-superset/releases
When we are able, this task is to upgrade our instance of Superset to 0.37.
Ref. T262162 for progress on the deployment project.
A new Superset release is available: https://github.com/apache/incubator-superset/releases
When we are able, this task is to upgrade our instance of Superset to 0.37.
Ref. T262162 for progress on the deployment project.
Too far into Q2 to address this. Pushing it until Q3 to avoid interfering with Q2 fundraising.
Hi everybody! Can I ask how is this instance working? Is it reading from Druid's analytics cluster? I am asking since I have some work planned in the backlog related to Superset (like removing old datasources not in sql alchemy, etc..) and I think it would be probably better to sync with you as the work proceeds. Moreover I can share more details about what changes (config wise etc..) with you between versions, so we can have the same baseline.
@elukey Howdy. This instance is fully self contained. We are currently reading only from our internal mysql datasources. You should feel free to do any datasource work you need as it won't affect us.
There is some desire to possibly query production data in the future, but that will require significant work on our end and is not currently on our roadmap.
Our deployment uses a significant puppet refactor since our puppet deployment is different. Otherwise, there is little difference in our base setup from production. Any info you want to share so that we keep a similar baseline would be greatly appreciated.
Reviewed the changes and tested the upgrade in virtualbox, it looks straightforward. The Superset install process includes a database update step, it's hard to tell whether what 0.37.2 does is incompatible with 0.36.0. Easy enough to back up the database first. Here's my recommendation for a procedure and rollback.
upgrade
rollback
This was a minor disaster. I tried upgrade using commit 7bdc41487a3, which is the version that was successful in virtualbox testing.
The upgrade fails during the db upgrade step, when it tries to create table that already exists:
sqlalchemy.exc.OperationalError: (MySQLdb._exceptions.OperationalError) (1050, "Table 'rls_filter_tables' already exists")
That was empty table, so I ended up working around the problem by rolling back the database, dropping the table, and starting over.