Second Grafana instance crashes after boot - failed to obtain lock

[danstadler-pdx]

Hi, I've been looking around for a docker-compose project that starts up Grafana in HA mode; several of them have had certain issues preventing from just starting up OOTB.

Your project gets pretty close, but the second Grafana instance crashes after boot. Here are the last 3 log lines from that container:

logger=sqlstore t=2025-07-30T18:37:43.342033808Z level=info msg="Connecting to DB" dbtype=postgres
logger=migrator t=2025-07-30T18:37:43.360496173Z level=info msg="Locking database"
logger=migrator t=2025-07-30T18:37:43.36327447Z level=error msg="Failed to lock database" error="failed to obtain lock"
Error: ✗ failed to obtain lock

Does this happen for you, and do you have an easy fix I can implement?

My setup:

• fork your project in GitHub
• open a Codespace on main
• run docker compose up -d

[danstadler-pdx]

I am trying a solution: in docker-compose, in the environment settings (for both Grafana service instances), I am adding this line:

- GF_DATABASE_MIGRATION_LOCKING=false

With that in place, all containers are starting correctly. Going to do some testing from this point.

I believe that the idea is that you set this value to false (on both services) on first launch, and then optionally again if you are performing an upgrade of Grafana. (In between, it's probably OK to set to true as an extra safeguarding mechanism, but I likely won't dig into that unless I find a specific reason to do so.)

The documentation for this config setting is here.