Bumping

I have the same behaviour, running supabase on docker inside a nested unprevilages LXC container on proxmox.

Should be fixed by #17587.

Edit the docker-compose.yml and update to the latest docker image which should be out next Tuesday, you can check here.

Please reopen and ping me if the issue still persists.

@ivasilov Problem Still exists with studio:v0.23.09_amd64 with this error:
Failed to load resource: the server responded with a status of 404 (Not Found)
:8000/api/integrations/default-org-slug?expand=true:1 Failed to load resource: the server responded with a status of 404 (Not Found)
:8000/api/integrations/default-org-slug?expand=true:1 Failed to load resource: the server responded with a status of 404 (Not Found)
localhost:8000/storage/v1/object/assets/IMG_6990 Large.jpeg:1 Failed to load resource: net::ERR_CONNECTION_REFUSED

Should be fixed by #17587.

Edit the docker-compose.yml and update to the latest docker image which should be out next Tuesday, you can check here.

Please reopen and ping me if the issue still persists.

@ivasilov, the latest image tag is v0.23.09

same issue as @aabulmagd.

any progress ?

@dfang having the same issue:

I've just updated the docker-compose with studio version v0.23.09 but still having those console log errors.

Same thing

same thing, the progress gets stuck at 0%.

any progress pls?

As it turns out I had a problem with the http (not https) specified in the .env, since I'm using nginx over kong with custom SSL certificates.

I managed to upload files with a python script.

Should be fixed by #17587.

Edit the docker-compose.yml and update to the latest docker image which should be out next Tuesday, you can check here.

Please reopen and ping me if the issue still persists.

Not help at all even update to latest docker image.

Video

Hi all — this still an issue? Just tested and looks resolved to me.

Got the same problem here on my side.

Running via Docker too and changed every localhost to my local ip (192.168.1.202) in the .env file. When uploading via bucket gui on my browser the POST method get postet to http://localhost:8000/storage/v1/object/files/

All GET methods are resolved under http://192.168.1.202:8000/api/storage/default/buckets

//edit

So today I got back to my office and tried to upload a file again. Now it routes correctly to my lan ip.
My .env values

SITE_URL=http://192.168.1.202:3000
ADDITIONAL_REDIRECT_URLS=
JWT_EXPIRY=3600
DISABLE_SIGNUP=false
API_EXTERNAL_URL=http://192.168.1.202:8000
SUPABASE_PUBLIC_URL=http://192.168.1.202:8000

Its strange it works today because i haven´t change anything overnight. Maybe its useful.

Got the same problem here on my side.

Running via Docker too and changed every localhost to my local ip (192.168.1.202) in the .env file. When uploading via bucket gui on my browser the POST method get postet to http://localhost:8000/storage/v1/object/files/

All GET methods are resolved under http://192.168.1.202:8000/api/storage/default/buckets

//edit

So today I got back to my office and tried to upload a file again. Now it routes correctly to my lan ip. My .env values

SITE_URL=http://192.168.1.202:3000 ADDITIONAL_REDIRECT_URLS= JWT_EXPIRY=3600 DISABLE_SIGNUP=false API_EXTERNAL_URL=http://192.168.1.202:8000 SUPABASE_PUBLIC_URL=http://192.168.1.202:8000

Its strange it works today because i haven´t change anything overnight. Maybe its useful.

It's worked for me. @supabase should auto modify script installation to detect localhost to IP.

Getting the same issue here, I am able to upload via SDK but not via studio GUI

I have the same error.

Hi guys, this still an issue? I've just pulled the latest version, and it works fine for me. If possible, could you guys post some images of the logs for storage or studio?

I have the same error when trying to upload. Can't pull the logs as when I run the default query I get:
{ "code": 502, "errors": [], "message": "Something went wrong! Unknown error. If this continues please contact support.", "status": "UNKNOWN" }

The DB seems to work fine. Via studio I add tables and data

The setup is brand new pull on hetzner cloud. No env changes as I understood from the docs that I should work fine as is (just testing here, so no security risks)

I've looked into this more closely, this comment solves it. You need to set SUPABASE_PUBLIC_URL to your URL in the .env file. It doesn't matter from which URL you're accessing the studio (you can even access it directly via its IP address), uploading of files (and some other minor features) are using that env variable since they're using the supabase-js library. You can see this behavior if you look closely in the URLs for the /upload POST and other GET requests.

If you're still seeing this issue, please share your SITE_URL, API_EXTERNAL_URL, SUPABASE_PUBLIC_URL and KONG_HTTP_PORT.

Still have the issue.

My env:
SITE_URL=http://49.***.***.68:3000/
API_EXTERNAL_URL=http://49.***.***.68:8000/
SUPABASE_PUBLIC_URL=http://49.***.***.68:8000/
KONG_HTTP_PORT=8000

it is a clean install on Hetzner I don't use for anything right now, so if you want to, you are welcome to investigate on the system itself.

it is a clean install on Hetzner I don't use for anything right now, so if you want to, you are welcome to investigate on the system itself.

Yeah, that would be great! Can you email me more info at [email protected] about your instance?

It reproduces when I use docker self-hosted deployment. There are 4 interfaces with 404 and the dashboard/projects cannot be accessed.

Had the exact same issue..
Changing all "localhost" to my IP in the .env file fixed it for me.

any update on this? Have the same issue

Same issue, changed all localhost to the appropriate LAN IP in the .env file. Docker restart. Still no uploads via studio.
I am using Nginx Proxy Manager. That's the only difference I can think of.

Ok, I got this working.
I'm accessing Studio externally. I'm using Nginx Proxy Manager.
I have a base url at 8000 and made a new url to point to 3000
In the .env file I changed:
SITE_URL=https://my-extradomain.com (no port - NPM does this for you)
and
API_EXTERNAL_URL=https://my-basedomain.com (again no port)
I also changed
SUPABASE_PUBLIC_URL=https://my-basedomain.com (no port)

Notice the https as it's a public route through your proxy.
I'm not a security expert so please feel free to let me know if this is not ok. It's working though. There are still lots of error in the javascript console but I think that's a different issue.

I'm having still those errors in console.log() but at functional level all seems to work correctly:

Ok, I got this working. I'm accessing Studio externally. I'm using Nginx Proxy Manager. I have a base url at 8000 and made a new url to point to 3000 In the .env file I changed: SITE_URL=https://my-extradomain.com (no port - NPM does this for you) and API_EXTERNAL_URL=https://my-basedomain.com (again no port) I also changed SUPABASE_PUBLIC_URL=https://my-basedomain.com (no port)

Notice the https as it's a public route through your proxy. I'm not a security expert so please feel free to let me know if this is not ok. It's working though. There are still lots of error in the javascript console but I think that's a different issue.

Just a quick intro to all these env vars:

• SITE_URL is the url to your Studio instance. Studio is optional and lives separate than the rest of the containers. This is the supabase.com/dashboard on our managed environment.
• API_EXTERNAL_URL and SUPABASE_PUBLIC_URL are the same thing and should point to where your db, auth, realtime and storage containers are. This is the same as <project-ref>.supabase.co on our managed environment. Studio also uses these env vars for some of the features.

I'm having still those errors in console.log() but at functional level all seems to work correctly:

Those are ok. Not important when self-hosting.

@ivasilov it would be cool to be able to remove the subscriptions module for self-hosted so to prevent all those errors.

• SITE_URL is the url to your Studio instance. Studio is optional and lives separate than the rest of the containers. This is the supabase.com/dashboard on our managed environment.
  This is my .env part with URLs:

## General
SITE_URL=http://localhost:3000
ADDITIONAL_REDIRECT_URLS=
JWT_EXPIRY=3600
DISABLE_SIGNUP=false
API_EXTERNAL_URL=http://localhost:8000

Can't open http://localhost:3000 in browser, http://localhost:8000 works well though. And also having a lot of 404 and not working AI SQL helper - trying to use gives a huge unclosable error popup (HTTP/1.1 500 Internal Server Error).

curl 'http://localhost:8000/api/ai/sql/edit' \
  -H 'Accept: application/json' \
  -H 'Accept-Language: ru-RU,ru;q=0.9,en-US;q=0.8,en;q=0.7,et;q=0.6' \
  -H 'Authorization: Basic YWRtaW46YWRtaW4=' \
  -H 'Connection: keep-alive' \
  -H 'Content-Type: application/json' \
  -H 'Cookie: _logflare_key=SFMyNTY.g3QAAAACbQAAAAtfY3NyZl90b2tlbm0AAAAYcWlXX2lhaFpudy1tOVMtR19GYnhpWmVEbQAAAAd1c2VyX2lkYQE.j8AUTkki4J8Fn9TAj7SLQkvMSUbyYEszcmwikTKltnY; __stripe_mid=120755d5-e809-4eb0-9064-c63216755c666c7008' \
  -H 'Origin: http://localhost:8000' \
  -H 'Referer: http://localhost:8000/project/default/sql/1' \
  -H 'Sec-Fetch-Dest: empty' \
  -H 'Sec-Fetch-Mode: cors' \
  -H 'Sec-Fetch-Site: same-origin' \
  -H 'User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36' \
  -H 'X-Request-Id: 7492f881-a6f6-4ecc-b373-f91b3668c586' \
  -H 'sec-ch-ua: "Chromium";v="122", "Not(A:Brand";v="24", "Google Chrome";v="122"' \
  -H 'sec-ch-ua-mobile: ?0' \
  -H 'sec-ch-ua-platform: "Windows"' \
  --data-raw $'{"prompt":"show current postgre version","sql":"select * from\\n  (select version()) as version,\\n  (select current_setting(\'server_version_num\')) as version_number;","entityDefinitions":["CREATE  TABLE public.mydata (\\n  id bigint GENERATED BY DEFAULT AS IDENTITY ,\\n  text text NOT NULL,\\n  CONSTRAINT mydata_pkey PRIMARY KEY (id)\\n) TABLESPACE pg_default;"]}'

Same issue on a fresh local supabase instance. I've tried to create a folder in my storage bucket and getting this error

_app-5e117e56c49dd8be.js:5 Refused to connect to 'http://127.0.0.1:54321/storage/v1/object/stickmonkey-media/aircrafts/.emptyFolderPlaceholder' because it violates the following Content Security Policy directive: "default-src 'self'  http://localhost:8000 http://localhost:8000 https://*.supabase.co wss://*.supabase.co https://*.hcaptcha.com https://cdn-global.configcat.com https://*.stripe.com https://*.stripe.network https://www.cloudflare.com https://1.1.1.1 https://*.vercel-insights.com https://api.github.com https://raw.githubusercontent.com". Note that 'connect-src' was not explicitly set, so 'default-src' is used as a fallback.
_app-5e117e56c49dd8be.js:5 Refused to connect to 'http://127.0.0.1:54321/storage/v1/object/stickmonkey-media/aircrafts/.emptyFolderPlaceholder' because it violates the document's Content Security Policy.

Ok, I got this working. I'm accessing Studio externally. I'm using Nginx Proxy Manager. I have a base url at 8000 and made a new url to point to 3000 In the .env file I changed: SITE_URL=https://my-extradomain.com (no port - NPM does this for you) and API_EXTERNAL_URL=https://my-basedomain.com (again no port) I also changed SUPABASE_PUBLIC_URL=https://my-basedomain.com (no port)

Notice the https as it's a public route through your proxy. I'm not a security expert so please feel free to let me know if this is not ok. It's working though. There are still lots of error in the javascript console but I think that's a different issue.

I am triy to answer this question. I'm also using Nginx Proxy Manager. When we need to access the storage service, Nginx Proxy Manager will proxy the request to our self-hosted server. And the configuration on the server shows that this request needs to be processed by localhost:8000. So, the javascript console shows that our request is forwarded to localhost:8000. And this localhost will be interpreted as the machine where our own client is located, not the localhost on the remote host. Therefore, the upload fails.

So, you change the service address in .env to your domain name (https://my-extradomain.com/). I think it works and is also secure. Because Nginx Proxy Manager will redirect to this address when making a request for the storage service. And this request is still SSL encrypted. It will not expose internal information of the server. How do others understand this process? Proxy Manager will redirect to this address when making the request to the storage service. And this request is still SSL encrypted. It will not expose the internal information of the server.

How do others understand this process?

Nginx Proxy Manager

AFAIK the connection between user and nginx will be secured (across the whole internet).

The connection between nginx and self-hosted supabase will be non-encrypted because nginx does SSL termination. In reality, this shouldn't matter because nginx and the supabase are usually on the same server OR in the same trusted virtual network.

I face the same issue: spinning loading bar at 0%.

The Console in Firefox first shows an issue with Cross-Origin and CORS and afterwards:

In Edge I get
POST http://localhost:8000/storage/v1/object/xyz/abc.jpg net::ERR_CONNECTION_REFUSED
and

I changed the ~/supabase/docker/.env from localhost to the real ip, as mentioned as a fix above. But it won't work and as you see, the POST request still uses localhost.

Can neither upload to buckets nor create users with authentication from the web UI

Fresh installed supabase sehfhost on VPS NGNIX (by https://blog.activeno.de/the-ultimate-supabase-self-hosting-guide). Same bug in studio - cannot add user nor create folder in the backet nor upload file.