Cloudflare

Yep! https://developers.cloudflare.com/r2/examples/aws/aws-sdk-js-v3/#generate-presigned-urls

Assuming I code this to support R2, and have all the variables in my .env file, would changing them to AWS S3 be all I need to migrate if needed? Also is it possible to trigger events based on R2 uploads?

Changing variables should ideally work to switch between R2 and S3

There isn’t a way to trigger events yet. If you want to do that, you’d have to do it from your app.

R2, S3

Is the cache limit of the Pro account a single file size not exceeding 512MB?

I have an MP3 file with a size of about 1G. It is reasonable to say that it should not be cached by CF, but I cannot play it directly when I access it, but must download it all. I turned off the cache rules and it was normal.

I turn off the caching rules and it will start playing directly instead of having to download the whole thing first

image

I'm following the example but I seem to get nested folders, is this intentional?

```ts
import { NextRequest, NextResponse } from 'next/server';

import { S3Client, PutObjectCommand } from '@aws-sdk/client-s3';
import { getSignedUrl } from '@aws-sdk/s3-request-presigner';

if (!process.env.S3_ENDPOINT) {
  throw new Error('S3_ENDPOINT is not defined');
}

if (!process.env.S3_ACCESS_KEY_ID) {
  throw new Error('S3_ACCESS_KEY_ID is not defined');
}

if (!process.env.S3_SECRET_ACCESS_KEY) {
  throw new Error('CLOUDFLARE_SECRET_ACCESS_KEY is not defined');
}

const S3 = new S3Client({
  region: 'auto',
  endpoint: process.env.S3_ENDPOINT,
  credentials: {
    accessKeyId: process.env.S3_ACCESS_KEY_ID,
    secretAccessKey: process.env.S3_SECRET_ACCESS_KEY,
  },
});

//  JSON body: { "filename": "my-file.txt" }
export async function POST(request: NextRequest) {
  const data = await request.json();
  const filename = data.filename;

  const url = await generateSignedUrl(filename);

  return NextResponse.json({ url });
}

async function generateSignedUrl(filename: string) {
  const command = new PutObjectCommand({
    Bucket: process.env.S3_BUCKET_NAME,
    Key: filename,
  });

  const url = await getSignedUrl(S3, command, {
    expiresIn: 60 * 60 * 24 * 7, // 7 days
  });
  return url;
}
```

I suppose you read Kian’s response, but yeah this is expected. There’s no such thing as “folders” really, the dashboard just uses / as a delimiter to fake them.

I'm a bit confused, when I then try to get the info of the file

```HEAD /camcorder-ai/406458e4-e71a-47de-800e-7ff910190619.mp4
NotFound: UnknownError```

```const headObjectCommand = new HeadObjectCommand({
        Bucket: process.env.S3_BUCKET_NAME,
        Key: parsedUrl.pathname,
      });

      const headObjectResponse = await S3.send(headObjectCommand);```

even though /camcorder-ai/file.mp4 is what I was returned back from the S3 command

I think the leading ‘/‘ is what your problem is

camcorder-ai/file.mp4 exists, but /camcorder-ai/file.mp4 does not.

I’m going off the dashboard screenshot above

I don't add the /camcorder-ai/ to any of my file upload code at all so I'm confused where it's coming from

How are you uploading these files? Some tools like to prefix object names with the bucket’s name.

Try setting the “forcePathStyle” key in your config to false?

didn't know that was a thing, will try it!

Yeah so S3 (and consequentially R2) has two ways of addressing buckets, one where the bucket name is in the hostname, bland another where the bucket name is in the path. 
Hostname style domains would look like: <bucketName>.<accountId>.r2.cloudflarestorage.com, and “path style” hostname would look like: <accountId>.r2.cloudflarestorage.com/<bucketName>. 
I bet you use the hostname variant as your S3_ENDPOINT env var, but S3 thinks you’re using the pathname variant (which is the default)

So it sends requests to <bucketName>.<accountId>.r2.cloudfalrestorage.com/<bucketName>/<objectKey>. That’s where the extra “camcorder-ai” is likely coming from. 

You can either set forcePathStyle to false, or switch your endpoint URL

```ts
const S3 = new S3Client({
  region: 'auto',
  endpoint: process.env.S3_ENDPOINT,
  credentials: {
    accessKeyId: process.env.S3_ACCESS_KEY_ID,
    secretAccessKey: process.env.S3_SECRET_ACCESS_KEY,
  },
  forcePathStyle: false,
});```

I set it to false but it still seems to do path style

```# These are the values for Cloudflare R2, they should be able to swapped to any other S3 compatible storage provider with minimal changes
S3_ENDPOINT='https://c38bc3761ea690e1a45693a40ce4fb2f.r2.cloudflarestorage.com/camcorder-ai'
S3_ACCESS_KEY_ID=
S3_SECRET_ACCESS_KEY=
S3_BUCKET_NAME='camcorder-ai'
S3_HOSTNAME="camcorder-ai.c38bc3761ea690e1a45693a40ce4fb2f.r2.cloudflarestorage.com"```

Remove the bucket name from there, the SDK will add it automatically

```
S3_ENDPOINT='https://c38bc3761ea690e1a45693a40ce4fb2f.r2.cloudflarestorage.com/'
S3_ACCESS_KEY_ID=
S3_SECRET_ACCESS_KEY=
S3_BUCKET_NAME='camcorder-ai'
S3_HOSTNAME="camcorder-ai.c38bc3761ea690e1a45693a40ce4fb2f.r2.cloudflarestorage.com"```

i was like I'm uploading files to root but a folder gets added

but then when I retrieve files I can't have the folder in the string!

Yeah haha unfortunately we can't make this "easier", this is just the S3 API contract

No probs! I think this should explain your 404s as well

perfect, I can now go through and remove the weird parsing I had to do on my keys to remove the "fake" folder

<@136415155016826880> 

<a:happydance:1071345120009539627>

i finally got something to upload to root

The actual filename for a given file would be `/foo/bar/baz` - 'folders' just delimit by `/`

If you want files to be in `/foo/bar/` then you'd just write to `/foo/bar/baz.jpg`

Hi! I am starting to use R2. I have few questions. I can easily download the files, but I can't upload/update one single file without deleting and reuploadind everything

Cloudflare is not caching a single file for my site

You're talking about a custom domain on your r2 bucket, I assume? Do you have a cache rule set up to force caching?

That sounds like something specific to the library/utility you are using

l have cache rules to cache images but l am not getting a single image cache

What's the rule you are using & the url of the image not being cached?

l am using R2 and getting millions of request and cloudflare doesn't cache me anything

That looks like a page rule, not a cache rule (although both should work). What's the URL that page rule is trying to match?

on other domain the cache is working fine

looks like one domain is having issue with cf

You probably want just `cdn.domain.com/*`
You might also want to move to Cache Rules at some point, they're easier overall and *eventually* Page rules will go away, but that shouldn't be the issue

You haven't show an actual link to an image or anything more, so it's hard to say much.
If you're just looking at the cache rate, that's not too helpful, it could be other items on your domain driving it down.

I would check the response headers on the image itself, looking for CF-Cache-Status.
There are other limits with Cache, the max file size that can be in cache for Free/Pro/Biz is 512mb

l got an email from CF part saying " we may decide to take action that could include: turning off cache to your domain(s),"