Restart loop: "Server failed to start" - error: getaddrinfo EAI_AGAIN voidauth-db

[Stephan-P]

I hope that someone can shine a light on my attempt to a first setup on an openmediavault 8 NAS (Linux 7.0.12+deb13-amd64), which gets stuck in a loop of voidauth container restarts.

This is the log shortly after docker UP:

voidauth_db  | 
voidauth_db  | PostgreSQL Database directory appears to contain a database; Skipping initialization
voidauth_db  | 
voidauth_db  | 2026-07-01 17:09:34.812 UTC [1] LOG:  starting PostgreSQL 18.4 (Debian 18.4-1.pgdg13+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 14.2.0-19) 14.2.0, 64-bit
voidauth_db  | 2026-07-01 17:09:34.812 UTC [1] LOG:  listening on IPv4 address "0.0.0.0", port 5432
voidauth_db  | 2026-07-01 17:09:34.812 UTC [1] LOG:  listening on IPv6 address "::", port 5432
voidauth_db  | 2026-07-01 17:09:34.813 UTC [1] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
voidauth_db  | 2026-07-01 17:09:34.817 UTC [32] LOG:  database system was shut down at 2026-07-01 17:06:32 UTC
voidauth_db  | 2026-07-01 17:09:34.820 UTC [1] LOG:  database system is ready to accept connections
voidauth     | 
voidauth     | 
voidauth     |  __   __ ______   __   _____    ______   __  __   ______  __  __   
voidauth     | /\ \ / //\  __ \ /\ \ /\  __ \ /\  __ \ /\ \/\ \ /\__  _\/\ \_\ \  
voidauth     | \ \ \ / \ \ \/\ \\ \ \\ \ \/\ \\ \  __ \\ \ \_\ \\/_/\ \/\ \  __ \ 
voidauth     |  \ \_/   \ \_____\\ \_\\ \_____\\ \_\ \_\\ \_____\  \ \_\ \ \_\ \_\
voidauth     |   \//     \/_____/ \/_/ \/_____/ \/_/\/_/ \/_____/   \/_/  \/_/\/_/
voidauth     | 
voidauth     |         ## Single Sign-On for Your Self-Hosted Universe ##
voidauth     | 
voidauth     | 
voidauth     | {"timestamp":1782925805872,"level":"info","message":"Starting VoidAuth Server"}
voidauth     | {"timestamp":1782925811054,"level":"error","message":"Server failed to start","errors":[{"name":"Error","message":"getaddrinfo EAI_AGAIN voidauth-db"}]}

Checking the log after a longer period shows an ongoing repetition of the VoidAuth header and the two last lines.

Remarkable: It takes unusually long for the database to get ready

My system runs Nginx Proxy Manager where I've created a Proxy Host for https://login.my-domain.tld incl. SSL-certificate termination, automatic https AND the Custom config as specified in the docs:

location / {
  include /config/nginx/snippets/proxy.conf;
  proxy_pass $forward_scheme://$server:$port;
}

My Compose (yes, docker Check confirms that the variable substitutions work):

services:
  db:
    image: postgres:18
    container_name: voidauth_db
    restart: unless-stopped
    volumes:
      - ./db:/var/lib/postgresql/18/docker
    networks:
      - voidauth-db
    environment:
      POSTGRES_PASSWORD: ${DB_PASSWORD}    # required, and must be the same as DB_PASSWORD in voidauth service
    healthcheck:
      test: "pg_isready -U postgres -h localhost"

  voidauth: 
    image: voidauth/voidauth:latest
    container_name: voidauth
    restart: unless-stopped
    ports:
      - '3000:3000'     # may not be needed, depending on reverse-proxy setup
      # - '3890:3890'   # needed if LDAP Server is enabled
    volumes:
      - ./config:/app/config
      # Only required for declaring OIDC Apps via docker labels (see OIDC-Setup documentation)
      # - /var/run/docker.sock:/var/run/docker.sock:ro
    networks:
      - voidauth-db
      - web-proxy
    environment:
      # Required environment variables, mostly obtained from file environment
      # See https://voidauth.app/#/Getting-Started?id=environment-variables for a list of possible environment variables
      APP_URL: ${APP_URL}
      STORAGE_KEY: ${STORAGE_KEY}          # required
      DB_ADAPTER: ${DB_ADAPTER}
      DB_HOST: ${DB_HOST}                  # required
      DB_PASSWORD: ${DB_PASSWORD}          # required, and must be the same as POSTGRES_PASSWORD in voidauth-db service
    depends_on:
      db:
        condition: service_healthy

networks:
  voidauth-db:
    name: voidauth-db
    driver: bridge
  web-proxy:
    external: true

Environment (sanitized):

# App settings
APP_URL=https://login.my-domain.tld
STORAGE_KEY=LONG_KEY1
STORAGE_KEY_SECUNDARY=LONG_KEY2
SESSION_DOMAIN=
DEFAULT_REDIRECT=
SIGNUP=false
SIGNUP_REQUIRES_APPROVAL=
EMAIL_VERIFICATION=true
MFA_REQUIRED=false
API_RATELIMIT=60               # 60 per second rate limit
ENABLE_DEBUG=false

# App customization
APP_TITLE=VoidAuth
APP_PORT=3000                  # listening port, in case the default 3000 needs to be changed
APP_COLOR=#906bc7              # default: #906bc7
APP_FONT=monospace             # default: monospace
CONTACT_EMAIL=

#Database settings
DB_ADAPTER=postgres            # default: postgres
DB_HOST=voidauth-db
DB_PORT=5432                   # default: 5432
DB_NAME=postgres               # default: postgres
DB_USER=postgres               # default: postgres
DB_PASSWORD=LONG_KEY3
DB_SSL=false                   # default: false
DB_SSL_VERIFICATION=true

#Database migration settings
MIGRATE_TO_DB_ADAPTER=postgres
MIGRATE_TO_DB_HOST=void-db
MIGRATE_TO_DB_PORT=5432
MIGRATE_TO_DB_NAME=postgres
MIGRATE_TO_DB_USER=postgres
MIGRATE_TO_DB_PASSWORD=
MIGRATE_TO_DB_SSL=false
MIGRATE_TO_DB_SSL_VERIFICATION=true

# SMTP settings
SMTP_HOST=mail.my-domain.tld
SMTP_FROM='VoidAuth<no-reply@my-domain.tld>'
SMTP_PORT=587
SMTP_USER=name@my-domain.tld
SMTP_PASS=password
SMTP_SECURE=false
SMTP_NOAUTH=false
SMTP_IGNORE_CERT=false
SMTP_TLS_CIPHERS=
SMTP_TLS_MIN_VERSION=

# LDAP settings
LDAP_ENABLED=false
LDAP_PORT=3890
LDAP_BASE_DN='cd=voidauth'
LDAP_BIND_DN='cn=ldap_bind,dc=voidauth'
LDAP_BIND_PASSWORD=
LDAP_TLS_CERT_FILE=
LDAP_TLS_KEY_FILE=

# Misc.
PASSWORD_STRENGTH=3
ADMIN_EMAILS=hourly
DEFAULT_USER_EXPIRES_IN=

[Stephan-P]

Reading various suggested solutions to this getaddrinfo EAI_AGAIN issue, I tried turning voidauth-db into a preexisting external bridge network. Alas, with the same result.

[Stephan-P]

Following this recommendation did not help

    ports:
      - 127.0.0.1:3000:3000       # may not be needed, depending on reverse-proxy setup
      - 192.168.2.211:3000:3000

[Stephan-P]

And commenting out the ports: array all together also doesn't bring resolve. Continued restarts of container voidauth with error "getaddrinfo EAI_AGAIN"

[Stephan-P]

Replacing all variable substitution in the compose file with direct values doesn't help either

[Stephan-P]

I could do with some expert help here. Anyone?

[Stephan-P]

One more thing that I could think of... the error message complains about not being able to get address information about voidauth-db
error: getaddrinfo EAI_AGAIN voidauth-db
But what is 'voidauth-db', the database or the bridge network?
So in order to avoid confusion, I changed the bridge network to 'voidauth-net' and the error remains: getaddrinfo EAI_AGAIN voidauth-db
So, the container 'voidauth' cannot obtain the address of the database 'voidauth-db'
Is there something that needs to be done with the database perhaps?

[Stephan-P]

Could ownership of the directory 'db' be an issue?

[notquitenothing]

Hello, the VoidAuth environment variable DB_HOST is the domain that VoidAuth should be able to use to reach the DB. In docker compose, the services in the same compose file can address each-other by their service name, and docker handles the DNS between them. Yours is not working probably because your postgres DB service is named db instead of voidauth-db. You can try to fix this issue by either renaming your postgres service to voidauth-db or changing your DB_HOST variable to DB_HOST=db. Let me know if that helps.

[Stephan-P]

Thanks for that pointer and explanation.
Got one significant step further.
Things to further look into:

• Somehow the URL doesn't work. What options should be enabled in Nginx Proxy Manager?
• SMTP-error. Check SMTP credentials, port, etc..

voidauth     |  __   __ ______   __   _____    ______   __  __   ______  __  __   
voidauth     | /\ \ / //\  __ \ /\ \ /\  __ \ /\  __ \ /\ \/\ \ /\__  _\/\ \_\ \  
voidauth     | \ \ \ / \ \ \/\ \\ \ \\ \ \/\ \\ \  __ \\ \ \_\ \\/_/\ \/\ \  __ \ 
voidauth     |  \ \_/   \ \_____\\ \_\\ \_____\\ \_\ \_\\ \_____\  \ \_\ \ \_\ \_\
voidauth     |   \//     \/_____/ \/_/ \/_____/ \/_/\/_/ \/_____/   \/_/  \/_/\/_/
voidauth     | 
voidauth     |         ## Single Sign-On for Your Self-Hosted Universe ##
voidauth     | 
voidauth     | 
voidauth     | {"timestamp":1783059733334,"level":"info","message":"Starting VoidAuth Server"}
voidauth     | {"timestamp":1783059733433,"level":"debug","message":"APP_URL: 'https://login.my-domain.tld'"}
voidauth     | {"timestamp":1783059733435,"level":"debug","message":"Session Domain: 'my-domain.tld'"}
voidauth     | {"timestamp":1783059733698,"level":"info","message":"Database schema updated."}
voidauth     | {"timestamp":1783059733699,"level":"info","message":"Connected to postgres database."}
voidauth     | {"timestamp":1783059734023,"level":"debug","message":"Attempting SMTP email connection."}
voidauth     | {"timestamp":1783059734332,"level":"info","message":"API and GUI listening on port: 3000"}
voidauth     | ====================================================================================================
voidauth     | 
voidauth     | Password reset link for the initial admin user, valid for 1 day:
voidauth     | 
voidauth     | https://login.mydomain.tld/reset_password?id=LONG_KEY1&challenge=LONG_KEY2
voidauth     | 
voidauth     | ====================================================================================================
voidauth     | {"timestamp":1783059735111,"level":"error","message":"SMTP email connection failed","errors":[{"name":"Error","message":"Invalid login: 535 Incorrect authentication data"}]}

[Stephan-P]

I decided to remove the custom config in Nginx Proxy Manager and now the VoidAuth Login page opens up alright.

[Stephan-P]

Did I misinterpret this section of the docs?

[Stephan-P]

The SMTP error was due to a silly typo in the credentials 😳

[Stephan-P]

I decided to remove the custom config in Nginx Proxy Manager and now the VoidAuth Login page opens up alright.

WRONG!
VoidAuth did open alright, but when it came to the real authentication stuff, it failed miserably with configuration errors.
For VoidAuth to do its thing in co-op with Nginx Proxy Manager, the NGINX Snippets need to be in place AND Custom config needs to be added to the Proxy Host configurations. See: Where do the NGINX Snippets go?