It’s a placeholder that isn’t recorded, it’s just there to support versioning *in the future*
It’s a placeholder that isn’t recorded, it’s just there to support versioning in the future
AAgain, this is also wrong, versions change when you replace an existing file.
AI suppose I could mimic object versions using the Cache API though.
ZWhat do you want the version information for? Only the latest upload under a key is stored.
Kthat isn’t recorded
KThere is no storage, there is no history, there is no rollbacks and there is no log.
KOnce you upload a new version, the string changes and you can do nothing with the old one.
KThe Cache API is ephemeral, per-colo and can be evicted at any time so it isn’t going to work for versioning.
AA specific use case, I upload an set of objects, and an index for a particular release that may occassionally change. On request for the initial index, I set a version cookie so all next requests will get the respective file, in case between requests the file was changed, to avoid mismatches.
ACache only needs to be accurate within the next few requests.
ZThat's not possible with R2. Like kian pointed out, the moment a new object is uploaded, the old one is completely gone
AAnd so, as I said, the cache API will suffice, or I could just add an arbitrary version to an objects key.
ZI would go with adding something to the key, since the cache API is in no way garunteed to stay around even for a single request
AIs there no guarentee at all?
Z...no, like I just said, none at all for the cache API
KIf you're relying on the only user accessing the old file is the exact same location that you uploaded the new file, it'll be a little more reliable - but that's about it.
KIf it's a separate connection, then it's probably hitting a new colo and therefore it won't be cached.
ARight, the goal is when they request an index, all next requests unless they request the same index (with a new value) will return the proper file version.
ASince in between the request, the index and contents potentially change.
AAStateful, in a way.
AI assume it should be fine, unless a user changes IP (and region) in the middle of requests?
KIt depends if this is critical or best effort.
KIt's probably fine, but you can run into issues where it's cached in Colo A but the user hit Colo B.
ACritical, it will break things if the wrong file is returned.
AIs there any documentation on Cache behaviours?
KEven in the same location & IP, you might hit a PoP with multiple colos and unless you have a connection kept alive (like in a browser), you have no particular affinity with a colo so it will hop around.
KIn a smaller location (i.e I'm hitting LHR with 6 colos), it's less of an issue.
AAnd this would cause a cache miss?
Sdoes accessing files in r2 through the cloudflare dashboard count towards total class A and B operations?
BHello, my system that was working yesterday started giving an error like this today, how can I solve it?
S3ServiceException [InternalError]: We encountered an internal error. Please try again.
S3ServiceException [InternalError]: We encountered an internal error. Please try again.
UThe S3 API seems fine for me. Can you try generating a new token and seeing if it works?
UOr just retrying the requests if you didnt. Sometimes errors happen 
Bi added my server's ip v6 address to filter and it worked 
Psorry guys if this was asked before a million times, but what desktop clients are compatible for browsing R2? Need it for debugging
KCyberduck works iirc
KSame with S3 Browser
KGenerally any client that supports S3 will work.
KSome don't, like Transmit by Panic
Ecyberduck may or may not mangle your content encoding though
K
SHow does one upload files to R2 via CLI?
HHello everyone, i have little problem about R2 CORS policy
HAnybody can help me?
HYou can only upload via Workers or S3 API
SDo they not have an easy-to-use CLI, through which I can just do
/cli upload-file <file_path> <destination_file_path>?Hmaybe this thing can help you https://www.npmjs.com/package/s3
npm
high level amazon s3 client. upload and download files and directories. Latest version: 4.4.0, last published: 8 years ago. Start using s3 in your project by running
npm i s3. There are 330 other projects in the npm registry using s3.Hhave basic schematic
Hnobody? 
