T
Twenty•3w ago
KrimmZZon

Unhealthy server node (redis connect fails) on 2GB instance

I have my node taking a lot of time running compose up - and end up failing due to server being 'unhealthy' After waiting for all the lines of logs for migration and cron jobs I see 
server-1  | /app/packages/twenty-server/dist/src/engine/core-modules/session-storage/session-storage.module-factory.js:57
server-1  |                     throw new Error(`Redis connection failed: ${err}`);
server-1  |                           ^
server-1  | 
server-1  | Error: Redis connection failed: Error: Connection timeout
server-1  |     at /app/packages/twenty-server/dist/src/engine/core-modules/session-storage/session-storage.module-factory.js:57:27
server-1  |     at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
server-1  | 
server-1  | Node.js v24.11.0
dependency failed to start: container twenty-server-1 is unhealthy
server-1  | /app/packages/twenty-server/dist/src/engine/core-modules/session-storage/session-storage.module-factory.js:57
server-1  |                     throw new Error(`Redis connection failed: ${err}`);
server-1  |                           ^
server-1  | 
server-1  | Error: Redis connection failed: Error: Connection timeout
server-1  |     at /app/packages/twenty-server/dist/src/engine/core-modules/session-storage/session-storage.module-factory.js:57:27
server-1  |     at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
server-1  | 
server-1  | Node.js v24.11.0
dependency failed to start: container twenty-server-1 is unhealthy
Not sure why :/ if I exec into redis container I can ping it (and this wouldnt start if that wasnt true) but on a 2GB node this keeps hitting redis connection failed. If I used 4GB it works without any changes. Docs say to have 2GB but it doesnt seem to be enough.
4 Replies
KrimmZZon
KrimmZZonOP•3w ago
In the past I have had it start randomly and then it works perfectly as I very rarely shut it down Also for my case can I set DISABLE_DB_MIGRATIONS and DISABLE_CRON_JOBS_REGISTRATION to true? Is it supposed to be done when the DB is directly on the bind mount?
Prastoin
Prastoin•3w ago
Hello there Hard to tell if it's an CPU issue here, also redis wouldn't be accessible or would shout OOM issues I would opt for a configuration issue are you sure that your server has access to your twenty instance directly ? If you're disabling auto migrations you will have to run them manually At what twenty version have you started your instance on ?
KrimmZZon
KrimmZZonOP•3w ago
Redis is pingable, but the twenty-server runs into Redis connection failed. This crashes the twenty-server and makes it restart then eventually leads to never starting My twenty-server is on the cloud instance yes, created from the docker compose pinned to 1.10.0 For automigrations - it takes a long while on the instance but is not the killer now as I increased the wait time.
Prastoin
Prastoin•3w ago
Are you able to ping your redis server from your twenty server instance though ssh? Might not be available from its scope

Did you find this page helpful?