https://discord.cloudflare.com logo
Join Discord
Powered by
# r2
  • j
    i think start-after and marker are semantically equivalent, no? big change is continuation-token, also owners should not be included in v2 results by default, and can be re-enabled with 
    fetch-owner
  • v
    I generally struggle to see how 
    continuation-token
    is really different from 
    marker
    aside from the fact that it's an opaque value (i.e. how would a hypothetical 
    NextStartAfter
    response value that you feed into 
    start-after
    be different from 
    NextContinuationToken
    ).
  • v
    i.e. I'm intellectually curious about why they split it up. The 
    fetch-owner
    thing I get since egress bytes are charged on S3 (but no one is complaining about the use of XML instead of a binary protocol or even JSON?)
  • e
  • e
    i should subscribe to that issue
  • j
    just tried cyberduck - you have to set a hidden pref to enable path style, and also anything over 100mb uses multipart upload, and fails. you can set another hidden pref to increase that 100mb threshold so it doesn't use multipart https://gist.github.com/johnspurlock-skymethod/6027c81953f81aa535d889d86a1061ba#cyberduck-cloud-storage-browser-for-mac-and-windows
  • e
    do you know how to write those prefs on cyberduck for windows?
  • j
  • e
    thanks!
  • j
    hmm, i'm not able to get synology cloud sync set up, chose "S3 Storage"-> "Custom Server URL", but no combo of settings there allows me to hit "next" and not receive "Authorization Failed". @itsmatteomanf do you have any info on what those settings looked like when you got it set up?
  • i
    Can you share a screenshot of what you set? I on my phone now, but I do remember how they look like.
  • i
    I’m gonna assume you did the correct things for the access keys (both id and token) given all the tests you did already.
  • j
    when the server address has the 
    https://
    prefix, it fails to list buckets properly, but without the prefix it's able to list buckets (!) (to populate the dropdown), but then hitting 'next' fails
  • j
    v4 vs v2 no difference
  • i
    That’s definitely missing the path-style routing… might be an older version of the package? 🤔
  • i
    That error seems like it’s the path-style thing.
  • i
    Lists the buckets, but goes no further.
  • j
    hmm just installed it
  • j
    the one from package center
  • j
    ooohhhh this is cloud sync, you mentioned hyper backup
  • j
    i don't see path-style as an option in hyper backup destination settings either, hmm...
  • j
    bummer
  • j
    and my hyper backup package version
  • j
    i bet it's because i'm not on dsm 7 yet
  • j
    whenever you get a chance, mind sending over the package versions of cloud sync and hyper backup you have on your side? maybe that setting is only avail on versions for dsm 7
  • e
    thanks for the documentation btw! with your settings, both Cyberduck and Mountain Duck work beautifully with R2 🧡
  • i
    Awesome! Thank you so much John and Matthew, working with MountainDuck for me too!
  • i
    Yeah, I'm on the latest DSM 7... Hyper Backup v3.0.2.
  • i
    One thing I noticed is that MountainDuck can't show upload progress for R2 uploads (just stuck on "Preparing" for a while, then completes suddenly). Is this an issue with how R2 is handling uploads or how MountainDuck works?
  • i
    In v3 is like this, only when you chose a Custom Server URL.
1...878889...1050Latest