Hi
I declare a error pod-template without a container named flink-main-container
to test rollback feature.
Please pay attention to the Pod-template in the old and new specs.
Last stable spec:
apiVersion: flink.apache.org/v1beta1
kind: FlinkDeployment
metadata:
name: basic-example
spec:
imag
It seems no other way to get the runtimeContext in a global commit. For me, I
think it's reasoable to propose the fetature.
I added flink-devs channel for more attention/discussion in flink devs.
Best regards,
Yuxia
- 原始邮件 -
发件人: "Tobias Fröhlich"
收件人: "User"
发送时间: 星期二, 2023年 2 月 14日
Hi, Xuekui.
As said in the exception stack, may be you can try to provide hint for the
function's parameters.
class ArrayTransformFunction extends ScalarFunction {
def eval ( @DataTypeHint("ARRAY") a: Array [Long],
@DataTypeHint("RAW") function: Long => Long): Array [Long] = {
a.map(e => f
Flink Cluster Context:
Flink Version - 1.15
Deployment Mode - Session
Number of Job Managers - 3 (HA)
Number of Task Managers - 1
Cancellation of Job fails due to following
org.apache.flink.runtime.rest.NotFoundException: Job
1cb2185d4d72c8c6f0a3a549d7de4ef0 not found
at
org.apache.flink.runti
Hi
I cannot see the difference between the two configurations, but the error
info `Failure executing: POST at: https://*/k8s/clusters/c-
fwkxh/apis/apps/v1/namespaces/test-flink/deployments. Message:
Deployment.apps "basic-example" is invalid` is strange. Maybe you can check
whether the configurat
Hi folks,
Is there a way we can configure the *dead letter queue* (dlq) for Kafka
source connector with *Table API? *Is Datastream API the only option for
now?
Thanks,
Eric
I make a test on the Application upgrade rollback feature, but this function
fails.The Flink application mode job cannot roll back to last stable spec.
As shown in the follow example, I declare a error pod-template without a
container named flink-main-container to test rollback feature.
However,
