@flamber Is this happening because of these warnings
2020-12-09 18:02:56,566 INFO metabase.util :: Maximum memory available to JVM: 1.0 GB
WARNING: sun.reflect.Reflection.getCallerClass is not supported. This will impact performance.
Warning: environ value jdk-11.0.7+10 for key :java-version has been overwritten with 11.0.7
Wow it’s confusing, when the log is in reverse order.
If you are starting multiple instances at the same time, then perhaps one of them has locked the database.
You are providing very little information about what you are doing. Are you upgrading, are you setting up a new setup, are you …?
@sagun.garg Seems like it’s dropping packets between Metabase and Postgres, so that’s the problem. I have no idea why it does that. I have not seen this problem before, but try restarting both RDS and ECS.
Still stuck at the same place(since last 2 days). I am using metabase/metabase:latest as my base docker image. Can you please recommend me another previous version image which is stable like metabase/metabase:v0.37.0 will that do any magic ?
@sagun.garg Try contacting AWS. I don’t know why you are seeing the error terminating connection due to administrator command, but there must clearly be a difference between your test setup and production.
@flamber Has anyone tested this latest docker image with AWS Fargate for a fresh self hosted solution with your own internal RDS Postgres DB version 10. This will help to confirm my 4th day struggle to install and self host in the first place.
@flamber I tried again with latest version image i.e v0.37.3 instead of v0.35.3 and I am stuck with memory allocation issues i.e max 2 GB allocation to JVM, hence I need help with latest stable image which is better than v0.35.3. Please advice
@sagun.garg You are misunderstanding the database connections.
Metabase has it’s own application database, where it stores all it’s information. That’s what you are setting with the environment variable.
Then you have data sources, which are the databases you want to view, and they are setup via the GUI.
After you setup a data source, then Metabase needs to sync+scan the database before you can use it in Metabase GUI - otherwise you will see the error Distinct, non-empty sequence of Field clauses https://github.com/metabase/metabase/issues/11354 - upvote by clicking on the first post
@flamber Thanks for that update, that literally threw me under the bus (but I managed to resolve this misunderstanding at my level) . I have sorted out the DB table issues as metabase level and data source level (separated the db concerns with two separate postgres DBs)
I ran the sync + scan
I am getting these errors in the Logs in Trouble Shooting Dashboard
**I tried the latest version on metabase/metabase:latest as well on AWS Fargate **
but then it gets stuck at the same JVM memory allocation issue + unrun database migrations,
please can you recommend an alternate predecessor image version number better than v0.35.3 and less than v0.37.3 (i.e latest version) which covers these sync errors and also manages to run on self-hosted AWS Fargate