Thanks for opening @RoadToDev101
This is a bug in the self-hosted docker-compose as the storage api seems not to listen on localhost, so it reports as unhealthy when it is fine.
This can be confirmed with docker logs $STORAGE_CONTAINER_ID

We will make a PR to address this. That said, the issue of not being able to create folders cannot be reproduced, can you provide logs from the storage container, please?

Hi @encima,

This is the log I get from the storage container when render the storage page on dashboard and create a new bucket name 'design_views'

{"level":30,"time":"2024-06-18T01:26:18.129Z","pid":1,"hostname":"a122cace4088","msg":"Server listening at http://0.0.0.0:5000"} {"level":30,"time":"2024-06-18T01:28:10.474Z","pid":1,"hostname":"a122cace4088","reqId":"req-1","tenantId":"stub","project":"stub","reqId":"req-1","type":"request","req":{"traceId":"req-1","method":"GET","url":"/bucket","headers":{"host":"storage:5000","x_forwarded_proto":"http","x_forwarded_host":"kong","x_forwarded_port":"8000","x_real_ip":"192.168.48.11","x_client_info":"supabase-js-node/2.41.1","accept":"*/*","user_agent":"node"},"hostname":"storage:5000","remoteAddress":"192.168.48.12","remotePort":43888},"res":{"statusCode":200,"headers":{"content_type":"application/json; charset=utf-8","content_length":"2"}},"responseTime":40.07513475418091,"msg":"stub | GET | 200 | 192.168.48.12 | req-1 | /bucket | node"} {"level":30,"time":"2024-06-18T01:28:26.236Z","pid":1,"hostname":"a122cace4088","reqId":"req-2","tenantId":"stub","project":"stub","reqId":"req-2","type":"request","req":{"traceId":"req-2","method":"POST","url":"/bucket","headers":{"host":"storage:5000","x_forwarded_proto":"http","x_forwarded_host":"kong","x_forwarded_port":"8000","x_real_ip":"192.168.48.11","content_length":"107","content_type":"application/json","x_client_info":"supabase-js-node/2.41.1","accept":"*/*","user_agent":"node"},"hostname":"storage:5000","remoteAddress":"192.168.48.12","remotePort":53754},"res":{"statusCode":200,"headers":{"content_type":"application/json; charset=utf-8","content_length":"23"}},"responseTime":131.50732326507568,"msg":"stub | POST | 200 | 192.168.48.12 | req-2 | /bucket | node"} {"level":30,"time":"2024-06-18T01:28:26.252Z","pid":1,"hostname":"a122cace4088","reqId":"req-3","tenantId":"stub","project":"stub","reqId":"req-3","type":"request","req":{"traceId":"req-3","method":"GET","url":"/bucket","headers":{"host":"storage:5000","x_forwarded_proto":"http","x_forwarded_host":"kong","x_forwarded_port":"8000","x_real_ip":"192.168.48.11","x_client_info":"supabase-js-node/2.41.1","accept":"*/*","user_agent":"node"},"hostname":"storage:5000","remoteAddress":"192.168.48.12","remotePort":53754},"res":{"statusCode":200,"headers":{"content_type":"application/json; charset=utf-8","content_length":"200"}},"responseTime":5.60968017578125,"msg":"stub | GET | 200 | 192.168.48.12 | req-3 | /bucket | node"} {"level":30,"time":"2024-06-18T01:28:26.349Z","pid":1,"hostname":"a122cace4088","reqId":"req-4","tenantId":"stub","project":"stub","reqId":"req-4","type":"request","req":{"traceId":"req-4","method":"POST","url":"/object/list/design_views","headers":{"host":"storage:5000","x_forwarded_proto":"http","x_forwarded_host":"kong","x_forwarded_port":"8000","x_real_ip":"192.168.48.11","content_length":"89","content_type":"application/json","x_client_info":"supabase-js-node/2.41.1","accept":"*/*","user_agent":"node"},"hostname":"storage:5000","remoteAddress":"192.168.48.12","remotePort":53754},"res":{"statusCode":200,"headers":{"content_type":"application/json; charset=utf-8","content_length":"2"}},"responseTime":7.82005500793457,"msg":"stub | POST | 200 | 192.168.48.12 | req-4 | /object/list/design_views | node"}

After that, I try to create a new folder in this bucket, the frontend shows that it created but when I reload, the folder is not there (I think it is not actually create the folder). And when I try uploading the file, it just spinning forever. And in these 2 actions, there is no log in the storage container.

Plus the studio container also unhealthy, but when check the log, it seems there is no problem. I hope this help.

thanks @RoadToDev101 , those logs show success for the storage requests so that certainly is weird!
Is there some issue with the volume mounting? Are you able to create a folder when you call the storage API directly?

Hi @encima,

I try to fetch the bucket api on Postman but I cant due to "Invalid authentication credentials". I try both Anon and Service Key for authorization. See the image below

The jwt in .env :
JWT_SECRET=girSzUKW1c937xjnJhKhuqLLM4WPIjNXrV8Odvra ANON_KEY=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.ewogICJyb2xlIjogImFub24iLAogICJpc3MiOiAic3VwYWJhc2UiLAogICJpYXQiOiAxNzE4NjQzNjAwLAogICJleHAiOiAxODc2NDEwMDAwCn0.STjbgWo03Cn8SRaRAn2bqw0M1io-k6g8GJkHAeUwXY4 SERVICE_ROLE_KEY=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.ewogICJyb2xlIjogInNlcnZpY2Vfcm9sZSIsCiAgImlzcyI6ICJzdXBhYmFzZSIsCiAgImlhdCI6IDE3MTg2NDM2MDAsCiAgImV4cCI6IDE4NzY0MTAwMDAKfQ.aq4Nv_JAsUJAJxwvr5Ro6GDyXVbyoGb9ij2UN40kJ_Y

Am I do it wrong?

@encima, so strange... I can upload the image through my NodeJS app.

@RoadToDev101 thanks for the update. so it works using the supabase client is that right? does it still fail with the dashboard?

from Postman, it looks like the credentials are probably failing due to them being incorrectly encoded or not picked up properly.

for supabase/studio, I think it is because of the starting period

I went inside the docker container and command

$ node -e "require('http').get('http://localhost:3000/api/profile', (r) => {if (r.statusCode !== 200) throw new Error(r.statusCode)})"

it gave me no error so i increased timeout in health check condition as below to fix unhealthy

supabase-studio:
  image: 'supabase/studio:20240422-5cf8f30'
  healthcheck:
    test:
      - CMD
      - node
      - '-e'
      - "require('http').get('http://localhost:3000/api/profile', (r) => {if (r.statusCode !== 200) throw new Error(r.statusCode)})"
    timeout: 15s
    interval: 5s
    retries: 5
  depends_on:
    supabase-analytics:
      condition: service_healthy

However it doesn't work with storage-api. I tried to debug the same way but got below result

/app # wget --no-verbose --tries=1 --spider http://localhost:5000/status
Connecting to localhost:5000 ([::1]:5000)
wget: can't connect to remote host: Connection refused

for supabase/studio, I think it is because of the starting period

I went inside the docker container and command

$ node -e "require('http').get('http://localhost:3000/api/profile', (r) => {if (r.statusCode !== 200) throw new Error(r.statusCode)})"

it gave me no error so i increased timeout in health check condition as below to fix unhealthy

supabase-studio:
  image: 'supabase/studio:20240422-5cf8f30'
  healthcheck:
    test:
      - CMD
      - node
      - '-e'
      - "require('http').get('http://localhost:3000/api/profile', (r) => {if (r.statusCode !== 200) throw new Error(r.statusCode)})"
    timeout: 15s
    interval: 5s
    retries: 5
  depends_on:
    supabase-analytics:
      condition: service_healthy

However it doesn't work with storage-api. I tried to debug the same way but got below result

/app # wget --no-verbose --tries=1 --spider http://localhost:5000/status
Connecting to localhost:5000 ([::1]:5000)
wget: can't connect to remote host: Connection refused

Are you connecting to storage from within the studio container? Either way, localhost will not work as it is not exposed there (unless on the storage container). You can use the same command but use studio instead of localhost

@encima I use the healthcheck command in the docker compose file for studio container and get this error:

@encima I use the healthcheck command in the docker compose file for studio container and get this error:

@RoadToDev101 that is expected yes. the studio is accessible through "studio" but not localhost.

I'm also facing the unhealthy issues on those two services. In addition, I also got the unauthorized issue when uploading files that are more than 6 MB; though, it works fine for those that are less than that.

Same issue here, the nginx server doesn't serve on localhost which makes the healthcheck provided in the github repo not working. Is there a workaround? Another healthcheck command. I tried to comeup with one, but without the ability to fetch from localhost it is kinda difficult to not hard code an IP.