Version 0.40 is here — let us know what you think

The new search functionality isn't very convenient, unfortunately. The idea is great, but it makes it hard to select a specific table from a specific database quickly, and that is what you need in 99% of cases. Perhaps, you can consider adding back the list filter on the database level.

Thanks!

@Alex There's a request for that:
https://github.com/metabase/metabase/issues/17011 - upvote by clicking :+1: on the first post

tried the custom deployment file which did get through the upgrade with no errors, but then it prompted me to recreate my account like i was a brand new user, which i've not seen before during any prior upgrade, so i reverted to 0.39.4

that's the second issue, which is that you have to decouple the database that was created with the Elastic Beanstalk deployment: https://www.metabase.com/docs/latest/operations-guide/creating-RDS-database-on-AWS.html#decouple-your-rds-database-from-the-elastic-beanstalk-deployment

so you are saying decouple rds from the elastic beanstalk deployment and then run the upgrade with the custom zip?

Exactly, or you can run the new zip (not the modified one) once you do the decoupling, so you can take advantage of Amazon Linux 2, the new Docker engine and upgrade from the old Elastic Beastalk platform (which you should do ASAP)

Excellent work! I'm using the jar file locally and on AWS EBS and this will help speed up the workflow in creating dashboards. I'd like to see a Treemap Visualization at some point. I work with agricultural data and the visual similarity to satellite imagery is intuitively understandable to farmers. Keep up the good work!

1 Like

Does this mean that there is not, and never will be, a way to update to 0.40+ without decoupling?

The issue is the following: if you create the database IN the Elastic Beanstalk bundle (the old bundle we had), it will create an RDS server with Magnetic storage (we tested it even with large resources and big disks, and it will always choose magnetic disks and not the provisioned IOPS disks as it should do for a production database. Why did AWS made this choice? we don't know) which is a very bad choice as it hurts the performance a lot. We want to ensure that our users have the best experience possible, so that's why we decided to teach our users how to decouple the database from the Elastic Beanstalk deployment (so our users can start a database with the optimal choices) and we strongly suggest to go this way.

It sounds like rather than "strongly suggesting" it is actually required. The performance was fine before. Sorry, but I'm pretty frustrated because I followed the decoupling guide five times, and rebuilt the EB environment five times, with no success. Each time I add the environment variables (and which ones I need to add is guesswork) it fails to apply the configuration. Updating used to be really easy but I've spent hours on this one with no success—for performance gains no one on my end is going to appreciate? :pensive:

try first decoupling your database, which is the first step. And once you have an RDS running which is not tied to Elastic Beanstalk, try deploying a new Metabase Elastic Beanstalk deployment from scratch BUT connect it to the old RDS, feel free to open a new thread with this issue, we'll be happy to help you.

1 Like

Any further updates on this? I've been holding off upgrading since it seemed like such a risky move. We're on 0.39.4 with the "old" platform and integrated rds instance.

What's the official way forward? Thanks.

We pushed a fix for old platforms and integrated rds instances. Be aware that AWS will stop supporting the old platforms next year. Always backup your database before upgrading

Understood. where can i get this fix? Thanks.

if you upgrade to the latest version now (40.3.1) it should have the fix. Once again, backup the database first

Hi there, I'm having some issues upgrading from 39.4 still. I have an integrated rds and using the elastic beanstalk deployment. On the upgrade it also asks me to setup from scratch. Is anybody else still having any issues with this?

Thanks.

Hi, if you are able to create a thread so we can see what's the error that your seeing on the upgrade that would be nice :slight_smile: