Re: flink-operator: jarURI seems mandatory

est, > Weihua > > > On Wed, Apr 5, 2023 at 5:51 PM David Causse wrote: > >> Hi, >> >> I'm trying to deploy a job (flink 1.16) with the flink-operator, the job >> jar is part of the image and placed under /opt/flink/usrlib. >> I thought that by placing the

flink-operator: jarURI seems mandatory

Hi, I'm trying to deploy a job (flink 1.16) with the flink-operator, the job jar is part of the image and placed under /opt/flink/usrlib. I thought that by placing the job jar there I could avoid setting the jarURI in the JobSpec but I'm getting a NPE (pasted at the end of this email) suggesting

Removing unused flink-avro causes savepoint to fail loading

Hi, While developing a job we mistakenly imported flink-avro as a dependency and then we did some cleanups. Sadly it seems that flink-avro has registered some kryo serializers that are now required to load the savepoints even though we do not use the functionalities offered by this module. The

Re: State processor API very slow reading a keyed state with RocksDB

oc.io/doc/org.rocksdb/rocksdbjni/6.20.3/org/rocksdb/ReadOptions.html#setFillCache(boolean) > -- > *From:* Seth Wiesman > *Sent:* Friday, September 10, 2021 0:58 > *To:* David Causse ; user > *Cc:* Piotr Nowojski > *Subject:* Re: State processor A

State processor API very slow reading a keyed state with RocksDB

Hi, I'm investigating why a job we use to inspect a flink state is a lot slower than the bootstrap job used to generate it. I use RocksdbDB with a simple keyed value state mapping a string key to a long value. Generating the bootstrap state from a CSV file with 100M entries takes a couple