cli: Supabase dev service no longer starting, vector container keeps restarting

Describe the bug Using supabase cli to do local dev. The start fails with the vector container restarting until healthcheck fails.

To Reproduce Steps to reproduce the behavior:

  1. run > supabase start … in project that was working
  2. starts until the vector container quits, starts, quits, starts, … health check fail
  3. failed running this also, still attempted health check > supabase start --ignore-health-check

Expected behavior Supabase services to start local like they used to.

Screenshots n/a

System information Rerun the failing command with --create-ticket flag.

  • Ticket ID: e8fd446f149a4b67aab8ce16d71549fa
  • Version of OS: Windows 11
  • Version of CLI: 1.183.5
  • Version of Docker: 27.0.3
  • Versions of services:
       SERVICE IMAGE      │      LOCAL       │ LINKED
 ─────────────────────────┼──────────────────┼─────────
   supabase/postgres      │ 15.1.1.61        │ -
   supabase/gotrue        │ v2.151.0         │ -
   postgrest/postgrest    │ v12.2.0          │ -
   supabase/realtime      │ v2.29.15         │ -
   supabase/storage-api   │ v1.0.6           │ -
   supabase/edge-runtime  │ v1.54.10         │ -
   supabase/studio        │ 20240701-05dfbec │ -
   supabase/postgres-meta │ v0.83.2          │ -
   supabase/logflare      │ 1.4.0            │ -
   supabase/supavisor     │ 1.1.56           │ -
   darthsim/imgproxy      │ v3.8.0           │ -

Additional context n/a

About this issue

  • Original URL
  • State: closed
  • Created 2 years ago
  • Reactions: 2
  • Comments: 17 (4 by maintainers)

Commits related to this issue

Most upvoted comments

I’ve tested a fix for this in cli 1.187.5. You will need to enable a few settings on docker desktop on windows (see screenshot).

  • expose daemon on tcp
  • allow *.docker.internal names

settings

After that, you can enable analytics without running into crashloop. The postgres and api gateway logs are still unavailable on studio but I will look into that separately. https://github.com/supabase/supabase/issues/26785

The workaround for now is to turn off analytics in config.toml

[analytics]
enabled = false

@sweatybridge that worked for me. I also enabled the settings in Docker. Thank you!

@sosensible the fix is only released to beta channel atm. Could you give npx supabase@beta start a try?

TLDR: only use alphanumeric for your project_id in config.toml

I had the same issue, tried upgrading to the beta version, tried analytics false which worked, but then had other issues, tried with health checks off, but then i noticed i had an odd project_id in my config.toml.

My project_id in my config.toml was project_id = "project-v-2.0.0"

I change my project_id to “project” and everything magically worked!

Some errors in docker from supabase vector were thread 'main' panicked at 'invalid authority: InvalidIpv4Address', src/sinks/util/uri.rs:122:69

My guess is the project_id, if it has too many periods in the text, causes some issue with the ip address when calling the analytics service.

supabase@beta works for me only when I set “analytics=false”, in this way vector container will not be started.

For us, supabase@beta is not working on Windows 11. Everything starts up, and with or without the ignore health check flag, it still shuts right down after everything starts up.

This should be reopened, it did not fix things. I created the setting above.

  • config.toml > analytics false
  • set the docker settings you outlined

The local supabase starts, and in less than a minute, after pulling current images. It is not showing the vector DB restarting, now it just fails the healthcheck perhaps, but the flag to ignore healthcheck had no impact. Still failed.

P.S. We used scoop to install supabase and are running on the current update.

I’m experiencing the same issue when trying to start it on Windows; however, it works fine on Mac.