Repository navigation
Open
Description
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
LxCunderProxmox VE - Operating system: Alpine
- Deployment:
Docker - Reverse proxy:
NginX Proxy Manageron anoterhLxCwithSSLenabled
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
Reacted by dows
Metadata
Metadata
Assignees
Labels
No labels
Activity
dows commentedon Feb 17, 2026
Same here. Sometimes it spins up after a couple redepolys and restarts
moonrailgun commentedon Mar 13, 2026
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:
You can also run
docker statswhile the container is starting to see if memory usage is hitting the limit.SKJoy commentedon Mar 18, 2026
You are absolutely correct. I ran a few tests and verified the following;
2 GBmemory;4 GBrecommended1 coreCPU;2 coreCPU recommendedI tried it on a Proxmox VE
LXCwith0.5 CPU limitand it failed, even with adequate memory!