https://discord.cloudflare.com logo
Join Discord
Powered by
# r2
  • k
    It does, since they're uploaded
  • k
    The checksum isn't an MD5 digest of the uploaded object though.
  • k
    It uses multipart by default for files above 16MB
  • k
    When I said the MD5 doesn't match the local one I meant this:
    Copy code
    csharp
                var getObjectMetadata = await s3client.GetObjectMetadataAsync(new GetObjectMetadataRequest()
                {
                    Key = fileKey,
                    BucketName = settings.CloudflareR2BucketName,
                });
                    
                if (getObjectMetadata.ETag == $"\"{md5}\"")
                {
                    Console.WriteLine($"File {fileName} already exists in the bucket with the same MD5 hash [{md5}].");
                    exists = true;
                }
  • k
    Is that incorrect?
  • k
    If I download the file from R2 it has the same MD5 has the local file
  • k
    If you're saying that the 
    ETag
    property is not the MD5 of the object, that is expected with multipart uploads.
  • k
    yeah, the etag isn't the MD5... https://stackoverflow.com/a/7347257/4744560
  • k
    do you have any idea on how to achieve this on multipart uploads?
  • n
    is there a difference between 
    <bucket>.<account-id>.r2.cloudflarestorage.com
    and 
    <account-id>.r2.cloudflarestorage.com/<bucket>
    , should i prefer one over the other?
  • k
    Nope, just different styles of addressing the same bucket
  • n
    okay, i thought the first got deprecated or something since i couldn't find that representation except in an older env variable
  • k
    Looks like I'm gonna be saving the MD5 checksum in the file's metadata
  • h
    What is the checksum based off of? I know when I rclone it can check against something, but is it using a different hash, or something else
  • k
    Apparently, on multi-part uploads the etag isn't the MD5 checksum: > ETag is a MD5. But for the Multipart uploaded files, the MD5 is computed from the concatenation of the MD5s of each uploaded part. So you don't need to compute the MD5 in the server. Just get the ETag and it's all. https://stackoverflow.com/a/61861691/4744560
  • h
    Interesting
  • k
    Yeah, I wasn't aware of that either xD
  • i
    Haha thanks for the explanation but it was a joke! Should’ve made that more clear. I well and truely agree that it’s the right decision to not include and delete all
  • a
    I'm trying to migrate some objects from other provider to R2, using rclone, but for some reason the API keeps returning me "Unauthorized" randomly
  • a
    Just running a simple list on the root directory is inconsistent
  • a
    Any idea on what could be causing this?
  • a
    Copy code
    2023/04/04 02:40:30 ERROR : : error listing: Unauthorized: Unauthorized
        status code: 401, request id: , host id:
2023/04/04 02:40:30 Failed to lsf with 2 errors: last error was: error in ListJSON: Unauthorized: Unauthorized
        status code: 401, request id: , host id:
  • a
    Well, I created a fresh pair of keys and the error went away
  • r
    Hey! I'm trying to upload a pdf to a cloudflare worker binded with R2 from my nextjs api backend... Do I need to use something like multer of formidable to handle multi part upload?
  • r
    ?r2-roadmap
  • h
    > 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)
  • c
    does each public bucket access counts as a class B operation
  • k
    If it's a cache miss, yes.
  • k
    Most requests from the same location to the same object, if at the same time, will be cached.
  • s
    am i dumb or why is my r2 bucket caching its 404?
1...975976977...1050Latest