JThank you!
ZI think you have to get enterprise support to get sla on r2
Tanyone know what could be causing this error?
im basically doing where base64FileData is a base64 encoded string. The transferring to buffer seems to be working fine, and using the same code to upload strings and the raw base64 seems to work but I just can't get it to work with the binary
im basically doing where base64FileData is a base64 encoded string. The transferring to buffer seems to be working fine, and using the same code to upload strings and the raw base64 seems to work but I just can't get it to work with the binary
TThe file is like 3MB btw
Tlol the I've been stuck on this for 2 days and the magic of posting something has worked again, I just have to upload slpFile.buffer instead of just slpFile
Tbad error message tho
AI would like to do some processing on the file uploaded to R2, however I was not able to decopress it ?
SUse list objects via any s3 wrapper
MHi!
I have files stored in R2, images and videos, and accessible through my domain
Is it possible to integrate cloudflare optimization images, so I can crop, transform and optimize it based on my url?
I have files stored in R2, images and videos, and accessible through my domain
Is it possible to integrate cloudflare optimization images, so I can crop, transform and optimize it based on my url?
Ryes, you can use the images product where you can just use the optimization
Rand it is very cheap
Ghello, can anyone help me using r2. because I just want to use it and want to learn it
II know it got blocked but just for future reference, there is over 25 thousand people on this server, so please don't try to ping 
@here 
Vdoes gcp need my account keys to migrate via super slurper?
Vor is it possible to have a private key on a specific bucket in gcp
Vcheers brah
BHey guys! How can I see my R2 objects that is uploaded locally in dev mode?
It seems that uploaded objects are stored in .wrangler/state/v3/r2/my-local-bucket/blobs/example-object.png.
Is it possible to see this image in browser like S3? Or it's not possible and should I create remote preview bucket in my account?
It seems that uploaded objects are stored in .wrangler/state/v3/r2/my-local-bucket/blobs/example-object.png.
Is it possible to see this image in browser like S3? Or it's not possible and should I create remote preview bucket in my account?
ZHello hello! I have a feature request...
TL;DR: CopyObject support for Sippy.
For context around my use-case: We are a music distributor and have a large-scale system which transcodes audio. Users upload music to us in various formats. We process it in various ways including transcoding the audio to an industry standard format and end up with a directory structure containing a bunch of XML files (containing metadata), images and audio. We then upload all this data in batches to ~30 partners around the globe (eg Spotify).
There are 2 costly parts to this. 1) Compute for the transcoding process. 2) Egress bandwidth.
We're running this system in GCP, where we store transcoded versions of audio in a 'cache' bucket so that we only need to transcode each audio file once. Due to the fact each of our partners wants batches in varying directory structures; we construct batches by copying the various files from GCS into a temporary 'batch build' bucket and then sync the data to our partners via SFTP with rclone. This relies on GCS cross-bucket copy operation.
We're migrating to R2; and want to use Sippy to pull data into our transcode cache bucket. However; when i'm trying to copy a not-yet-in-R2 file from our transcode bucket to our batch build bucket CopyObject returns a not found error.
So.. if CopyObject was 'sippy aware' and could trigger the pull down of data into the source bucket on request; THEN copy it.. that would be the ideal behaviour. In the meantime my workaround is to first check for the existence of an object in my source bucket and issue a get request if it's not been migrated yet... then the copy operation works.
Hope this makes sense!
TL;DR: CopyObject support for Sippy.
For context around my use-case: We are a music distributor and have a large-scale system which transcodes audio. Users upload music to us in various formats. We process it in various ways including transcoding the audio to an industry standard format and end up with a directory structure containing a bunch of XML files (containing metadata), images and audio. We then upload all this data in batches to ~30 partners around the globe (eg Spotify).
There are 2 costly parts to this. 1) Compute for the transcoding process. 2) Egress bandwidth.
We're running this system in GCP, where we store transcoded versions of audio in a 'cache' bucket so that we only need to transcode each audio file once. Due to the fact each of our partners wants batches in varying directory structures; we construct batches by copying the various files from GCS into a temporary 'batch build' bucket and then sync the data to our partners via SFTP with rclone. This relies on GCS cross-bucket copy operation.
We're migrating to R2; and want to use Sippy to pull data into our transcode cache bucket. However; when i'm trying to copy a not-yet-in-R2 file from our transcode bucket to our batch build bucket CopyObject returns a not found error.
So.. if CopyObject was 'sippy aware' and could trigger the pull down of data into the source bucket on request; THEN copy it.. that would be the ideal behaviour. In the meantime my workaround is to first check for the existence of an object in my source bucket and issue a get request if it's not been migrated yet... then the copy operation works.
Hope this makes sense!
TCan you find somewhere the expected cost of what you used since the last billing date or something?
TWow weird that that doesn't exist
TTNice
Jpricing question: R2 Class A Operations are $4.50 / million requests - does it round up to the nearest million for billing?
CI thought it didn't do that before 
I know Cache Reserve was that way though
I know Cache Reserve was that way though
Cehh nvm I might have just never passed through the free Class A ops and I'm thinking about the storage costs
Uhmm the pricing for R2 doesnt say it rounds thats kinda messed up
U$9 for being 8k over 1 million is rough
Jyea its... surprising (s3 does not do this) - perhaps someone orange can weigh in on if this is official? if true, seems like something that would be useful to document: https://developers.cloudflare.com/r2/pricing/
UI wonder how many products on cf do this now 
JDurable Objects Compute Duration definately does
Jit's been undocumented for a while? : )
Usince when has it ever been prepaid? :wat:
UIve never prepaid for it always post
Jyea perhaps the examples on the pricing page would at least implicitly reflect this if all of the example amounts weren't evenly divisible by 1,000,000
Jstill think it's worth a mention - as other cloud paygo platforms are not like this
USo if its prepaid then and you paid for 2 million that means on the next month any unused in that 2nd million wont be billed?
UThen its not prepaid
UIts postpaid and rounded
Uand not documented
UIt is though the bill he got was 1.008m and it rounded up to 2m
ULiterally every platform on earth charges by each single request (for the most part). So if they are rounding up to the nearest million it should be documented
UI mean every other S3 provider also shows the request price by millions by charges per request so why would anyone think it rounds up to the next million?
UThat doesnt make any sense
UI guess technically aws and gcp show by thousand (not million) but charge per