Please unpublish this ASAP, this template does not use a volume and will cause data loss for anyone who uses it.

Added a volume, better now?

Nope, I don't know where seaweed stores uploaded files but I'm sure it's not /tmp

I appreciate you wanting to get into publishing a template, but if it's going to come at the cost of unknowing users loosing their data, it should not be published.

Additionally, this template comes with no authentication what so ever.

Please unpublish, you are welcome to republish when you are more familiar with seaweed and Railway templates.

I just tried it and it works:
1. Deploying template and attaching public URL to map to port 8333 (seaweedfs-production-60d4.up.railway.app)
2. Locally, I created a dummy file: echo 123 > /tmp/testfile

3. Uploaded it to seaweed: aws --endpoint-url https://seaweedfs-production-60d4.up.railway.app s3 cp /tmp/testfile s3://testbucket/testfile

4. Checked it is there: aws --endpoint-url https://seaweedfs-production-60d4.up.railway.app s3 ls testbucket
   Got: 2024-10-02 10:59:05 4 testfile

5. You can check my test project, project id: 2e3d00a0-12dd-475e-9115-a42ce7a69b68
   
   PS, I know you probably didn't mean this, but I have to say you really made me feel like the work I put into this wasn't worth it - your second comment was much better btw, maybe trying to work together to fix these things would be a better approach with new users.
   
   Regarding auth, in my use case I don't need auth as it will be used internally only in a private network, I guess users can add their own auth on top.

> I just tried it and it works

Yes you can upload files just fine, but if you upload a file and then redeploy that file will disappear since seaweed is not putting files into /tmp if seaweed doesn't store files in the volume you will not have any persistence.

> PS, I know you probably didn't mean this, but I have to say you really made me feel like the work I put into this wasn't worth it

I understand that I could have come off as rude, but given this template was just a single service with no special work done I was focused on thinking about the end users who could lose data, I don't like telling users that we don't restore data.

> maybe trying to work together to fix these things would be a better approach with new users.

Happy to work with you on templates, feel free to come to the community Discord where I can help, much better approach than publishing a semi-broken template.

Ok I think I fixed it, valid points, thanks for discussing it.

Changed mount to /data
Changed the deploy to weed server -s3 -dir /data

Tested the file I created is still there after a redeploy.
Better?

PS, I think y'all should add the license of these templates too, AGPL is a no-go for many organizations.

Better!

Now I understand you may not need auth, but most users would want auth, I think you should look into adding that.