Issues with Docker container on port 5432

Hi,

I'm having an issue running PostgresML in a Docker container, on an Ubuntu 22.04 server.
Like this

docker run -d --name postgresml --restart always  \
    -v postgresml_data:/var/lib/postgresql \
    -p 5433:5432 \
    -p 8089:8000 \
    ghcr.io/postgresml/postgresml:2.8.2 

The Ubuntu 22.04 server is running Postgres 16 locally on port 5432 (not on a Docker container).
Hence my exposing the containerized Postgresml on port 5433.
So I don't see any reasons why I should be getting the errors below when I try to run the container

Starting PostgresML

• Starting PostgreSQL 15 database server
  ...done.
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server
• Removed stale pid file.
  Error: /usr/lib/postgresql/15/bin/pg_ctl /usr/lib/postgresql/15/bin/pg_ctl start -D /var/lib/postgresql/15/main -l /var/log/postgresql/postgresql-15-main.log -s -o -c config_file="/etc/postgresql/15/main/postgresql.conf" exited with status 1:
  2024-03-29 16:16:22.637 UTC [21] LOG: starting PostgreSQL 15.6 (Ubuntu 15.6-1.pgdg22.04+1) on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0, 64-bit
  2024-03-29 16:16:22.637 UTC [21] LOG: listening on IPv4 address "0.0.0.0", port 5432
  2024-03-29 16:16:22.643 UTC [21] FATAL: lock file "/var/run/postgresql/.s.PGSQL.5432.lock" already exists
  2024-03-29 16:16:22.643 UTC [21] HINT: Is another postmaster (PID 22) using socket file "/var/run/postgresql/.s.PGSQL.5432"?
  2024-03-29 16:16:22.648 UTC [21] LOG: database system is shut down
  pg_ctl: could not start server
  Examine the log output.
  ...fail!
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server
  ...done.
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server

Any idea why this is happening and how to fix it?

Thanks

[NyakudyaA]

To debug this better try the following

• remove the volume declaration from the run command and let the container use an internal volume. If this works then the issue is with the volume
• use a different volume when mapping I.e pg-data, you could try a local volume VS docker volume

[astrooom]

I'm facing exactly the issue described currently. I have tried both a docker volume:

volumes:
  - postgresml_volume:/var/lib/postgresql

And no declared mount (letting the container use an internal volume). Both give the same output as in #1392 (comment)

I've tried both ghcr.io/postgresml/postgresml:2.9.2 and ghcr.io/postgresml/postgresml:2.7.13 with the same result.

Additionally, I'm not running another postgresql instance on the same port. This indicates that there is something wrong within the container itself.

EDIT: After a couple restarts and rebuilds, its now showing these logs:

Starting PostgresML

• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]
  Starting dashboard
  Starting PostgresML
• Starting PostgreSQL 15 database server [ OK ]

Not sure what changed exactly. The container keeps restarting though. It might be the same issue just not logged properly.

[SilasMarvin]

Sorry for the delayed response. Do you have any updates on this? Is this still an issue or is there any other information you can share?

[stephenljk]

just delete the lock file, i removed mine and restart the container, its running fine after remove.

Sorry for the delayed response. Do you have any updates on this? Is this still an issue or is there any other information you can share?

[maxifelix]

Hi,
I ran into this issue twice, when the host server that runs the docker container restarts.
I couldn't find a definitive solution.
So deleting the lock file is the easiest course of action.
This is the command I used, with the postgresml running:

docker exec -it [container_name_or_id] rm /var/run/postgresql/.s.PGSQL.5432.lock
I hope it helps someone.