Feedback on Kubernetes self hosted Install of preview version

I wanted to share here all the tricks I had to perform in order to switch my self hosted version (to avoid others to face same problems)

I was running the 8.0.2.b37747 version with LDAP

but switch to 9.0.0-beta.b42121 version was unsuccessful. Redeploy of the same version was failing.

So I give a try to preview version from an empty environment.

  • No issues with install & upgrade commands
  • Failures at run with exceptions with postgres connection: solved by adding the rule: SQLALCHEMY_ENABLE_POOL_PRE_PING: “true” (read here)
  • once the applicaton stable I’ve applied the backup of the DB
  • I had to manage the Postgres schema update on notification_destinations table (I removed the only entry I had from my backup)

New interface is great, and fullscreen button really useful.

2 bugs (but I don’t know if already raised),

  • the list of favorites dashboard / query is not completed: It seems that it only displays the stars of the first fetch and not the full list. I can raise an issue with screenshots if needed

  • the schema from the query creation page is endless spinning and does not answer anything (PG or Kusto data sources)

Technical details:

  • Redash Version: 9.0.0-beta.b42121 (28/01/21)
  • Browser/OS: Chrome v87
  • How did you install Redash: Kubernetes 1.19.4 self hosted
2 Likes

You may not need/want it, but there is a Helm chart here that has worked well for me.

Thanks for your answer;
I’m using the same one.

It is now working fine if I add the queues
QUEUES: "queries,celery,schemas,periodic,default"
to the adhocWorker

My bad, I’ve missed that Pull Request: WIP: Changes required for redash v9 by jimsparkman · Pull Request #36 · getredash/contrib-helm-chart · GitHub

Once the rebase done, it solved all my issues;

It creates generic Workers & Schedulers that are processing the schemas & periodic queues