playing audio via R2 is so smooth

Some operations occur when you open the dashboard

Current limitations

Location Hints are only honored the first time a bucket with a given name is created. If you delete and recreate a bucket with the same name, the original bucket’s location will be used.

Are you guys planning to fix this limitation when deleting and recreating the bucket eventually? That would be really nice

Is it possible for me to create an API token granting access to a single bucket for R2? Right now it seems anyone who has access to any of my R2 tokens can access all buckets. This seems very dangerous.

Not possible currently, but coming soon

Hi, I have a bunch of geojsons (roughly 300k) I would want to serve from a R2 bucket, what would be the best way to go about uploading them on R2 gzipped (or brotli'd if that's even possible) and have R2 return them correctly to a browser client? I seem to be having trouble with uploading through rclone as the R2 only responds with the incorrect content encoding/type

It'll do that natively as long as you're uploading them with

Content-Encoding: gzip

Based on the client's

Accept-Encoding

header, it will be decompressed on-the-fly to plain-text or served as-is in gzip.

I'm guessing I also have to pass the content-type on upload, right?

Recommended, but not related to GZIP

That's just good practice so browsers know what to do (download or render in tab, etc)

Thanks, I'll try it out

So, in case anybody stumbles in the same trouble as me down the line ``` rclone copy --progress --header-upload "Content-Type: application/json" --header-upload "Content-Encoding: gzip" 2023-03-27 r2inat:inat-geojsons/2023-03-27```is the way to go. Using

--header

over

--header-upload

results in authorization errors with rclone, files are now served correctly with the accept header from the request

thank you

good morning prestiges, i'm quite sad not finding how to solve an issue i have with r2

i want to share file in r2 with an expiring link

but i prefer bucket level presigned urls

which if i am currect is a buck-based-presigned-link/file-name.ext

my idea is to generate the bucket link and fetch it when i need to get a file

Sounds like you just want a normal presigned URL?

so i will save these presigned link and generate new one if it expires

okay, how do i go about it?

and i'd be happy if you explain to me how presigned URL works so i will better know how to solve the issue

They’re implemented the same as S3 - https://developers.cloudflare.com/r2/api/s3/presigned-urls/

i have stumbled upon all docs available but still can't understand how to resolve this issue

i only need the presigned to download this files not for uploading them

What is R2 sending me?

Should I stop downloading now?

Where are you seeing that?

I download a zip file via

Cyberduck

on MacOS Monterey 😐