https://discord.cloudflare.com logo
Join Discord
Powered by
# r2
  • m
    wrangler r2 bucket delete $bucket --force
    does not seem to work either. Do I have to create a program to iterate through the bucket´s object in order to delete the bucket?
  • k
    The AWS CLI is a separate tool from the APIs, 
    --recursive
    is something they made themselves, it just iterates your bucket objects and removes them all, the APIs don't add support for it
  • m
    No, true. But list objects allow for prefix searching, which is what I see it´s not working in R2´s API.
  • k
    I haven't used the AWS CLI with R2, but I know that 
    rclone purge
    will properly remove everything and then the bucket. Make sure to give it more worker/scanner processes
  • m
    Thank you karew!
  • k
    Do you know approximately how many items are in the bucket?
  • m
    1.8M
  • k
    This should work:
    Copy code
    rclone purge --progress --verbose --transfers=40 --checkers=40 --max-backlog=-1 --order-by=name your-remote:your-bucket
  • k
    Ideally you run this on a sever that has rclone installed instead of a local computer
  • m
    Working nicely, thank you so much @Karew !
  • s
    You should be able to get a pretty accurate estimate if you look at your bucket's stats (on the bucket's page). Class A/B ops and storage are all that's billed according to https://developers.cloudflare.com/r2/pricing/
  • s
    The source (Wasabi/R2) shouldn't affect your cache hit rates. Are you usage patterns the same? Are you requesting larger files now?
  • l
    THanks sid
  • p
    Nope. No change on usage pattern. Everything is the same after switching. It's been 24 hours and the cache hit rate is still at 20 - 30 %.
  • k
    Tiered Cache does not work with R2, it only uses the regular cache
  • k
    Which could be the issue
  • p
    Aahh! Yes, that's what I thought!
  • p
    the data reflects that.
  • p
    is there any reason why that is the case? I thought it should work well with tiered caching
  • k
    It would be wonderful if it did. There is some sort of architectural issue with R2 that prevents it right now. (R2 is secretly a Worker, which could be related)
  • p
    ok that make sense. Thank you Karew! Let's hope they make it work together one day!
  • k
    If your cache was previously 80%, it should recover more, but slower. As long as files are being accessed a lot from a particular region.
  • k
    I think Tiered Cache used to be paid product? I am also not sure why (1) we need to turn on Tiered Cache (2) it isn't just the way the entire CDN works no matter what
  • p
    Yes that seems to be the case. It was around 10% when I switched, it's not 35% after 24 hours. I'll stick with R2 for a few days and see how it goes :). R2 seems to deliver faster than wasabi with tier caching in most cases.
  • p
    Yeah, it was bundled with Argo but split out on its own.
  • h
    It may have something to do with the fact that it affects cache invalidations(i.e. invalidating an asset at the origin may take a bit longer to propagate).
  • p
    I can't wait for R2 to come with cross region replication. With 2 buckets serving in North America and Asia, that would be perfect for websites with a global audience.
  • a
    does R2 have jurisdiction support yet?
  • s
    Not yet, but it's being actively worked on, so it should land soon-ish
  • a
    how much will i pay for 20TB network usage of get requests from 2gb storage a month?
1...103810391040...1050Latest