Tianji Docker instance failing to start service

[SKJoy]

Tianji won't start

Been trying for days, with latest and other prior releases, with completely fresh instances, and severly failed to bring Tianji up and running.

I will miss Tianji for sure :(

Container log loop

> tianji@1.31.13 start:docker:db /app/tianji
> cd src/server && pnpm db:migrate:apply && pnpm ch:migrate:apply
2026/02/14 12:17:07 Send request error: Post "http://localhost:12345/serverStatus/report": dial tcp 127.0.0.1:12345: connect: connection refused
> @tianji/server@ db:migrate:apply /app/tianji/src/server
> prisma migrate deploy
2026/02/14 12:17:12 Send request error: Post "http://localhost:12345/serverStatus/report": dial tcp 127.0.0.1:12345: connect: connection refused
Prisma schema loaded from prisma/schema.prisma
Datasource "db": PostgreSQL database "tianji", schema "public" at "tianji_postgres:5432"
2026/02/14 12:17:17 Send request error: Post "http://localhost:12345/serverStatus/report": dial tcp 127.0.0.1:12345: connect: connection refused
79 migrations found in prisma/migrations
No pending migrations to apply.
2026/02/14 12:17:22 Send request error: Post "http://localhost:12345/serverStatus/report": dial tcp 127.0.0.1:12345: connect: connection refused
> tianji@1.31.13 start:docker /app/tianji
> pnpm start:docker:db && pnpm start
> tianji@1.31.13 start:docker:db /app/tianji
> cd src/server && pnpm db:migrate:apply && pnpm ch:migrate:apply
> @tianji/server@ db:migrate:apply /app/tianji/src/server
> prisma migrate deploy
2026/02/14 12:17:35 Start reporting...
2026/02/14 12:17:35 Mode: http
2026/02/14 12:17:35 Version: 1.0.0
2026/02/14 12:17:40 Send request error: Post "http://localhost:12345/serverStatus/report": dial tcp 127.0.0.1:12345: connect: connection refused

Environment

• Infrastructure: Linux LxC under Proxmox VE
• Operating system: Alpine
• Deployment: Docker
• Reverse proxy: NginX Proxy Manager on anoterh LxC with SSL enabled

Docker compose

  tianji-redis:
    image: redis:latest
    restart: unless-stopped
    volumes:
      - ./Tianji/Redis/Data:/data
      - ./Tianji/Redis/Configuration:/usr/local/etc/redis

  tianji_postgres:
    image: postgres:15.4-alpine
    restart: unless-stopped
    environment:
      POSTGRES_DB: tianji
      POSTGRES_USER: tianji
      POSTGRES_PASSWORD: tianji
    volumes:
      - ./Tianji/Postgres:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U $${POSTGRES_USER} -d $${POSTGRES_DB}"]
      interval: 5s
      timeout: 5s
      retries: 5

  tianji:
    image: moonrailgun/tianji:latest
    restart: unless-stopped
    tmpfs:
      - /tmp
    depends_on:
      - tianji-redis
      - tianji_postgres
    environment:
      PUBLIC_URL: ""
      ALLOW_REGISTER: false
      EMAIL_SERVER: "** REDACTED **"
      EMAIL_FROM: "** REDACTED **"
      JWT_SECRET: "** REDACTED **"
      CUSTOM_TRACKER_SCRIPT_NAME: "tianji.js"
      AMAP_TOKEN: "** REDACTED **"
      MAPBOX_TOKEN: "** REDACTED **"
      AUTH_CUSTOM_TYPE: "oidc"
      AUTH_CUSTOM_NAME: "** REDACTED **"
      AUTH_CUSTOM_ISSUER: "** REDACTED **"
      AUTH_CUSTOM_ID: "** REDACTED **"
      AUTH_CUSTOM_SECRET: "** REDACTED **"
      AUTH_GITHUB_ID: ""
      AUTH_GITHUB_SECRET: ""
      AUTH_GOOGLE_ID: ""
      AUTH_GOOGLE_SECRET: ""
      DISABLE_AUTO_CLEAR: false
      DISABLE_ANONYMOUS_TELEMETRY: true
      ALLOW_OPENAPI: true
      ENABLE_FUNCTION_WORKER: true
      DATABASE_URL: postgresql://tianji:tianji@tianji_postgres:5432/tianji
      REDIS_URL: redis://tianji-redis:6379
    ports:
      - 0.0.0.0:12345:12345

[dows]

Same here. Sometimes it spins up after a couple redepolys and restarts

[moonrailgun]

Hi, thanks for reporting this.

From your logs, the database migration completes fine, but the server process seems to get killed silently right after — no error message at all. This usually means the system ran out of memory (OOM).

Tianji needs more memory during the first startup because it runs database migrations. Once the server is up and running, it uses much less. We recommend at least 2 GB RAM for the initial startup.

If your server doesn't have that much physical memory, you can add swap space as a workaround:

sudo fallocate -l 2G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile

You can also run docker stats while the container is starting to see if memory usage is hitting the limit.