yexinru/doc-exports
1
0
Fork 0
forked from docs/doc-exports
Code Pull Requests Activity
Files
a41a4e0331fbc0beff63bb28fa59228cc9a751df
doc-exports/docs/dli/umn/dli_03_0121.html
Su, Xiaomeng 3a07c40729 dli_umn_20250212 
Reviewed-by: Pruthi, Vineet <vineet.pruthi@t-systems.com>
Co-authored-by: Su, Xiaomeng <suxiaomeng1@huawei.com>
Co-committed-by: Su, Xiaomeng <suxiaomeng1@huawei.com>
2025-02-24 08:11:08 +00:00

1.7 KiB
Raw Blame History

How Do I Configure Connection Retries for Kafka Sink If it is Disconnected?

Symptom

You used Flink 1.10 to run a Flink Opensource SQL job. The job failed after the following error was reported when Flink Sink wrote data to Kafka.

Caused by: org.apache.kafka.common.errors.NetworkException: The server disconnected before a response was received.

Possible Causes

The CPU usage is too high. As a result, the network is intermittently disconnected.

Solution

Add connector.properties.retries=5 to the SQL statement.

create table kafka_sink(
     car_type string
    , car_name string
    , primary key (union_id) not enforced
) with (
    "connector.type" = "upsert-kafka",
    "connector.version" = "0.11",
    "connector.properties.bootstrap.servers" = "xxxx:9092",
    "connector.topic" = "kafka_car_topic ",
    "connector.sink.ignore-retraction" = "true",
    "connector.properties.retries" = "5",
    "format.type" = "json"
);
Parent topic: Flink SQL Jobs