hello, has anyone faced the issue?

Cannot create recoverable writer due to This s3 file system implementation does not support recoverable writers., will use the ordinary writer.

as far as I can see, flink has successfully created a checkpoint, but these logs still appear. i am using presto and s3p:// format and have copied the Presto JAR to the plugins and lib directories of my Flink deployment.

Only

flink-s3-fs-hadoop

is supported for file sinks

hmmm just to clarify, in my case i don`t have explicit file sink i have kafka source and kafka sink i use filesystem as checkpoin-storage and use s3 and presto to store checkpoints

ok, I apologize then, usually that exception happens when using a file sink

You shouldn’t add it to your lib folder

@sap1ens no problem, it sounds better, cause i have much more troubles in case of using hadoop)

And make sure that you don’t have a Hadoop S3 jar still somewhere 😅

yes, it was in test purposes, previously it was only in /plugins, but i asked for help from chatGPT and this is the advice)

ChatGPT definitely doesn’t take the latest Flink changes into account 😅

100%) as i remember, his knowledge limited by 2020 but he is still pretty useful)