Did someone mention Cyberduck before?

How do you actually use that to manage files though?

There is visual clients that you can use with S3, like S3 Browser or Cyberduck

If you don't need a graphical user interface, there is CLI options like rclone that work well

I was just talking about this in another thread

their "enterprise" plan only scare away and make them sell less

really? I would pay 2k for "enterprise" just to get the few specific features I want if it was just a button to click & pay

does this apply to workers too?

no, Workers always run

You can use cache in Workers, but your Worker will still always be invocated / you pay for the request

and we asked you in the other channel as well to stop it - Cloudflare has a flow to sign up for Ent, just like most other companies

Enterprise is highly customizable, you get billed for everything from requests/bandwidth/etc, and you ask for specific features (stream, waf ml, zero trust features, etc). Most of the limits, like max cache size, websockets, can all be raised/changed in your contract., can't really be click and pay

you'd get more $$ by charging x for the "enterprise" features and then letting everyone know about it and giving a button to pay 🤷‍♂️

We've already told you in the past to drop the conversation about this. Would be greatly appreciated if you just stopped complaining about it - its not changing.

Ok, so I have discovered that whatever algorithm R2 is doing for ETags on multipart uploads, it's different from what S3 does (see vlovich's comment on https://github.com/s3tools/s3cmd/issues/1273). I still can't find any documentation for what algorithm it does use. I want to verify the CRC MD5 of a local file matches the remote file, and without that info I seem to be unable to use multipart uploads and that makes me sad.

I don’t think it’s deterministic

Can't you provide your own etag?

FWIW talks of making the Etags deterministic have come up. (May not necessarily be MD5 like S3 though) For now you’ll have to use other ways to do that, but I think it should still be possible?

I guess you could store it as custom metadata if you control what response header your tooling will be using - otherwise I suspect the usual checksum request headers are purely response-time validation rather than being stored with your object.

so list_objects_v2 won't return the metadata though? I now need to do a loop over every object to get the metadata?

it still seems like if you want to verify local files and not totally throttle downloads, you basically have no choice but to use non-multipart uploads with R2.

If it helps to add some context - in my simple test case, a small (4GB) game server upload takes about 77 seconds on my fiber with multipart enabled (about 400mbit/s overall), and about 353s without (89mbit/s). I've optimized the non-multipart path to use parallel workers who are each doing uploads, and starting with the largest files first, but in the end there's one especially big file that ends up being the real bottleneck. I'm still not really sure why boto3 isn't faster at uploads in general. So if I'm patching a game to fix of a major player issue, this means the game will be broken for an extra 4 minutes while we wait for the upload without multipart. I expect my lovely 4GB server build will only grow from here, and probably be 7-15GB, so this issue could get worse, not better over time.

i generated a presigned url to get a file from cloudflare r2. and now i want to make download link of that file. i did Download btn. but when you click on the link it redirects to cloudflare link. download doesnt start. anyone help me it is urgent for me.

If you mean it’s just opening a tab and displaying the file in the browser, you should be setting this header: https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Disposition

i setted the Content-Disposition header in cors then how to make it download. normally we do download. what to do now?

Content-Disposition: attachment

or

Content-Disposition: attachment; filename="example.jpg"

where to use this when generating the presigned url or when downloading from frontend?

my frontend is nextjs and backend is django. i added the content disposition while generating presigned link in backnd. i fetched the link in frontend using api. now how to download in frontend?

How do I link AWS S3 to my Cloudflare R2 buckets so I can upload files bigger than 300 MB and use FTP (at no additional cost if possible)?