R2 only currently supports migrating the entire bucket

Hi, is there a scoped role for giving a user access only to R2?

Is it possible to get R2 object public URL via the API? Did not find it in the HEAD request.

You can connect multiple subdomains or r2.dev to your bucket, the API doesn't return full URLs. You just need to concatenate

https://<subdomain>/<object-key>

Hey I am having issue with R2 downloading large files that there could be transfer interruption around 1 TB

Tried wget and curl none of them worked

For large files, it might be beneficial to use Range requests to download the file in chunks.

chunking doesn't work for me because I am using pipeline untar command like

wget -qO- ... | tar -xv

Not sure what else you can do. As with uploads, we recommend downloading in chunks, to hedge against network issues. Even if your ISP & Cloudflare are performing perfectly, there is always a chance that there is a hiccup in some intermediary, and thus it is a lot easier to recover from that if you download in smaller bites. Maybe you could try downloading the chunks manually, and then run it through tar?

Problem is I don't have a large enough disk it's a 3 TB file and my drive is like 3.5 TB lol

This should apply with any storage provider, not just Cloudflare

s3 always worked without issue though

🤷

?r2-roadmap

> These features are now available!! > - **Object Lifecycles**: > - Public buckets, with custom domains: > - Presigned URLs in the S3-compatible API > > Future roadmap > Read more: > - Jurisdictional Restrictions (e.g. 'EU') > - Live Migration without Downtime (S3->R2)

Hi, I would like to try R2 Migrator, but there is one thing: The cloud storage bucket you are migrating consists primarily of objects less than 10 GB (1000³ bytes). Objects greater than 10 GB will be skipped and need to be copied separately. Does it mean, that one object cant be greater that 10GB or all objects together?

No single object may be greater than 10 GB

First, I would like to use the R2 as a backup storage for my current S3 storage (DO spaces). Is it possible to set up worker to sync R2 with my current S3 storage ? Or only solution is small VPS with rclone ?

Hi folks. Is the any plan to support 404 pages from R2 public buckets with custom domains soon ? Thanks!

Hi, I'm trying to configure public access to my R2 bucket with a custom domain. I recently transferred this domain over to Cloudflare, and it is now added as one of my sites on Cloudflare and everything seems in order. However, when I try to connect the domain to my R2 bucket, I keep getting this error. DNS record for this domain already exists on zone. (Code: 10056)

Hi all. I'm looking into enabling users to create buckets that only they can access. I was thinking of doing this via the binding to workers. Is that a reasonable way to go about this? Looks like you might need to know the name of a bucket before binding a bucket to a worker? So I'm not sure if there's a way to dynamically bind buckets to workers, or if you would have to create a new worker per bucket. Or maybe I'll have to use presigned urls or some other alternative mechanism for this?

You already have a record set up for whatever subdomain you’re trying to use. For example, if you’re trying to use

files.example.com

, you need to remove the CNAME or whatever you might have set for

files

already

Is there any chance to increase the bucket limit per account? I'd like to use presigned Urls for users to edit a set of objects and I'd rather have many buckets containing only a few objects than a few buckets with huge amounts of objects.

Nvm I figured it out! Thanks

I think I've encounted a bug in R2. When I use the

CopyObjectCommand

to copy an object, and I attempt to set

ContentDisposition

on the new object as a part of the

CopyObjectCommand

, the new object exists, but the

ContentDisposition

is not set. This is my copy code:

ts
  await client.send(
    new CopyObjectCommand({
      Bucket: bucket,
      CopySource: `${bucket}/${tmp}`,
      Key: key,
      ContentDisposition: contentDisposition,
    }),
  );

The new object at

key

exists, but when I use

HeadObjectCommand

I don't see a content disposition. This is the reponse:

json
{
  "$metadata": {
    "httpStatusCode": 200,
    "attempts": 1,
    "totalRetryDelay": 0
  },
  "AcceptRanges": "bytes",
  "LastModified": "2023-04-29T02:42:09.000Z",
  "ContentLength": 4662718,
  "ETag": "\"a8b4bcf0648ec79e2e79cb116fda9e7c\"",
  "ContentType": "video/mp4",
  "Metadata": {}
}

The AWS S3 documentation suggests that

ContentDisposition

should be returned: https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/preview/client/s3/command/HeadObjectCommand/ I've confirmed through debugging that the value for

ContentDisposition

is a string as such:

attachment; filename=This%20is%20an%20oink%20moo.mp4;

You need to specify

MetadataDirective: "REPLACE"

and set all the metadata to the new values it should be

hi, i was wondering why the public url only supports http/1.1

http/2 would be quite beneficial for my use case

R2 public domains are HTTP2 now, but if you happen to have one that was created a long time ago, you may have to remove and re-add it to get HTTP2