404's are cached by default for... 5 minutes if I recall correctly

ah i see

so even if somethings uploaded it'll still have a 404?

Until the cache clears, in that location, yes.

until that cache value is expired

gotcha

ah yeah i see it expired after probably about 5 minutes

thanks

it’d be nice if the custom domain for public bucket don’t have to be of the same account

Not sure if this is a worker/cache question or r2 question. Ive got a worker serving and caching some images from r2 with the file path like

/avatars/<userid>.jpg

. I noticed that if i replace that file at the path with the same key in r2, the previously cached image is still served sometimes. Is that expected behavior? I thought it would return the updated file since the object's

etag

would be different.

It'll never check the new ETag since it's cached

Only when it expires or it's a cache miss will it have the new one

Cache is per-colo so it could be cached in one colo and not another

thanks! that makes sense I was trying to get around doing some mass updates in my Firestore DB . Im storing the avatar urls in a few different spots throughout

Does Kyber PQ play any role in R2? Other CF products have used Kyber interested to know if there are any plans to use it for encryption or why it might not be a good solution.

what is wrong with rclone? 3KiB/s 😦

i'm using rclone copy ~/Dev/md r2prod:md --progress

Try restarting the transfer

did so many times

but same thing

What’s your tested speed on that server?

is it just me or is r2 currently terrible slow

i was looking atmigrating my current bucket from gcp cloud storage and jeez..

i am seeing 200ms-400ms response times for lots of my uploads

not sure if i am uploading things wrong or what

Doesn't look like it to me. One of my apps is getting (mostly) consistent sub-100ms Range requests.

How big are your files?

not super big, i mostly have lots of json/image files