[Bug] [Flink] After K8s Session runs the job, the console is always in the state of executing the job.

[13535048320]

Search before asking

• I had searched in the issues and found no similar issues.

What happened

K8s Session 运行作业后控制台一直处于执行作业的状态，任务提交成功的，Flink UI 里能看到任务正常运行，运维中心看不到对应的实例，或者处于 Unknown 状态，特别是在启动的作业多了之后，会频繁出现这种情况，没有报错日志

版本：
Dinky 1.0.1
Mariadb 10.6.11
Paimon 0.8
Flink 1.17.2
Kafka 3.4

FLINK SQL:

set 'execution.checkpointing.interval' = '30s';
set 'execution.checkpointing.max-concurrent-checkpoints' = '1';
set 'execution.checkpointing.tolerable-failed-checkpoints' = '1';
set 'execution.checkpointing.min-pause' = '30s';
set 'table.exec.state.ttl' = '45min';
set 'execution.checkpointing.timeout' = '1h';
set 'table.exec.sink.upsert-materialize' = 'NONE';

CREATE CATALOG paimon WITH (
 'type' = 'paimon',
 'warehouse' = 's3://uat-warehouse/paimon',
 's3.endpoint' = 'http://minio:9000',
 's3.access-key' = '',
 's3.secret-key' = '',
 'fs.s3a.connection.maximum'='2000',
 'fs.s3a.threads.max'='4000',
 'fs.s3a.buffer.dir'='/tmp/'
);

use catalog default_catalog;
create database if not exists source_kafka ;

set 'parallelism.default' = '12' ;
create table default_catalog.source_kafka.source_s4h_kafka_qm_zqminspectdefect
(
 `CHARG` STRING,`POSNR` STRING,`FEGRP` STRING,`FECOD` STRING,`ANZFEHLER` DOUBLE,`FEHLBEWC` STRING,`PRUEFLINR` STRING,`LASTRECTXNTYPE` STRING,`RECCREDATE` STRING,`LASTRECTXNDATE` STRING,`LASTRECTXNTIME` STRING,`LASTRECTXNUSERID` STRING
) with (
'connector' = 'kafka'
, 'topic' = 'ZQMINSPECTDEFECT.A20.SAP'
, 'properties.group.id' = 'source_s4h_qm_zqminspectdefect_sink'
, 'scan.startup.mode' = 'group-offsets'
, 'properties.bootstrap.servers' = 'kafka.confluent:9071'
, 'properties.auto.offset.reset' = 'earliest'
, 'properties.security.protocol' = 'SASL_SSL'
, 'properties.sasl.jaas.config' = 'org.apache.kafka.common.security.plain.PlainLoginModule required username="" password="";'
, 'properties.sasl.mechanism' = 'PLAIN'
, 'format' = 'json'
, 'json.fail-on-missing-field' = 'false'
, 'json.ignore-parse-errors' = 'true'
);

insert into paimon.ods.ods_s4h_qm_zqminspectdefect/*+ OPTIONS('sink.use-managed-memory-allocator'='true', 'sink.managed.writer-buffer-memory'='256M') */
select 
 `CHARG` ,`POSNR` ,`FEGRP` ,`FECOD` ,`ANZFEHLER` ,`FEHLBEWC` ,`PRUEFLINR` ,`LASTRECTXNTYPE` ,`RECCREDATE` ,`LASTRECTXNDATE` ,`LASTRECTXNTIME` ,`LASTRECTXNUSERID` ,
 cast( FROM_UNIXTIME(UNIX_TIMESTAMP()) as TIMESTAMP(3) ) as data_update_time 
from default_catalog.source_kafka.source_s4h_kafka_qm_zqminspectdefect ;

What you expected to happen

作业正常启动，运维中心状态正常

How to reproduce

1. 新建 K8s Session 集群配置，jobmanager.flink:8081
2. 新建flink sql任务，选择 K8s Session 模式
3. 运行flink sql多个项目

Anything else

No response

Version

1.0.0

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[github-actions]

Hello @13535048320, this issue is about K8S, so I assign it to @gaoyan1998 and @zackyoungh. If you have any questions, you can comment and reply.

你好 @13535048320, 这个 issue 是关于 K8S 的，所以我把它分配给了 @gaoyan1998 和 @zackyoungh。如有任何问题，可以评论回复。

[gaoyan1998]

可以发一下日志吗

[13535048320]

@gaoyan1998
这两天的日志
dinky-2024-04-11.log
dinky-ws-2024-04-10-1.log

[gaoyan1998]

重启一下dinky，就可以了，这个会在1.0.2修复