Re: Trouble sinking to Kafka

Hi Marco, The documentation kind of suggestion this is the cause: > https://nightlies.apache.org/flink/flink-docs-release-1.12/dev/connectors/kafka.html > > However, I think the documentation could benefit with a few examples and > scenarios that can ill-considered configurations. > Matthias alre

RE: Trouble sinking to Kafka

Good morning Marco, Your fix is pretty plausible: * Kafka transactions get started at the beginning of a checkpoint period and contain all events collected through this period, * At the end of the checkpoint period the associated transaction is committed and concurrently the transactio

Re: Restart from checkpoint - kinesis consumer

Thanks Danny, Let me comeback with results. > > On Feb 23, 2022, at 3:41 AM, Danny Cranmer wrote: > >  > Hello Vijay, > > > Once i do that my flink consumer need to be restarted with changed > > parallelism. > Why is this? The Flink consumer continuously scans for new shards, and will > aut

[Flink-1.14.3] Restart of pod due to duplicatejob submission

Hello, Recently due to log4j vulnerabilities, we have upgraded to Apache Flink 1.14.3. What we observed we are getting following exception, and because of it pod gets in crashloopback. We have seen this issues esp. during the time of upgrade or deployment time when existing pod is already running.

Re: Flink SQL: HOW to define `ANY` subtype in constructured Constructured Data Types(MAP, ARRAY...)

There is no built-in LogicType for ’ANY’, it’s a invalid token > 2022年2月23日 下午10:29，zhouhaifengmath 写道： > > > When I define a udf paramters like: > public @DataTypeHint("Row") Row > eval(@DataTypeHint("MAP") Map mapData) > > It gives error: > Please check for implementation mistak

Re: Trouble sinking to Kafka

I fixed this, but I'm not 100% sure why. Here is my theory: My checkpoint interval is one minute, and the minimum pause interval is also one minute. My transaction timeout time is also one minute. I think the checkpoint causes Flink to hold the transaction open for one minute, and thus it times o

Re: Flink job recovery after task manager failure

Hi, Afek! When a TaskManager is killed, JobManager will not be acknowledged until a heartbeat timeout happens. Currently, the default value of heartbeat.timeout is 50 seconds [1]. That's why it takes more than 30 seconds for Flink to trigger a failover. If you'd like to shorten the time a failover

Re: Trouble sinking to Kafka

Hi Marco, I'm no expert on the Kafka producer, but I will try to help. [1] seems to have a decent explanation of possible error causes for the error you encountered. Which leads me to two questions: if (druidProducerTransactionMaxTimeoutMs > 0) { > > properties.setProperty("transaction.max

Flink job recovery after task manager failure

Hi, I am trying to use Flink checkpoints solution in order to support task manager recovery. I’m running flink using beam with filesystem storage and the following parameters: checkpointingInterval=3 checkpointingMode=EXACTLY_ONCE. What I see is that if I kill a task manager pod, it takes f

Flink SQL: HOW to define `ANY` subtype in constructured Constructured Data Types(MAP, ARRAY...)

When I define a udf paramters like:    public @DataTypeHint("Row") Row eval(@DataTypeHint("MAP")     Map mapData)It gives error:    Please check for implementation mistakes and/or provide a corresponding hint.        at org.apac

Re: Task manager errors with Flink ZooKeeper High Availability

Thanks for the response on this issue. with the same configuration defined *high-availability: zookeeper* *high-availability.zookeeper.quorum: zk-noa-edge-infra:2181* *high-availability.zookeeper.path.root: /flink* *high-availability.storageDir: /flink_state* *high-availabi

Re: Restart from checkpoint - kinesis consumer

Hello Vijay, > Once i do that my flink consumer need to be restarted with changed parallelism. Why is this? The Flink consumer continuously scans for new shards, and will auto scale up/down the number of shard consumer threads to accommodate Kinesis resharding. Flink job/operator parallelism does

Restart from checkpoint - kinesis consumer

Hi Team, I am running flink 1.11 kinesis consumer with say N kinesis shards, but i want to increase/decrease shards to N+M or N-M. Once i do that my flink consumer need to be restarted with changed parallelism. But i am unable to restart from existing checkpoint because of change in number of s

Re: java.lang.Exception: Job leader for job id 0efd8681eda64b072b72baef58722bc0 lost leadership.

Hi Nico, Thanks for getting back to us. We are using Flink 1.14.0 and we are using RocksDB. We currently are using the default memory settings. We'll look into increasing our managed memory fraction to 0.6 and see what happens. Do writes to ValueStates/MapStates have a direct on churn of the F

Re: java.lang.Exception: Job leader for job id 0efd8681eda64b072b72baef58722bc0 lost leadership.

Hi Jai, On Tue, Feb 22, 2022 at 9:19 PM Jai Patel wrote: > It seems like the errors are similar to those discussed here: > - https://issues.apache.org/jira/browse/FLINK-14316 > - https://cdmana.com/2020/11/20201116104527255b.html > I couldn't find any other existing issue apart from the one you

Re: CSV join in batch mode

Hi, Killian Sorry for responding late! I think there is no simple way that could catch csv processing errors. That means that you need to do it yourself.(Correct me if I am missing something). I think you could use RockDB State Backend[1], which would spill data to disk. [1] https://nightlies.apac