You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "sigalit (Jira)" <ji...@apache.org> on 2025/01/26 16:20:00 UTC

[jira] [Created] (FLINK-37228) kubernetes HA fails both on flink 1.19.1 and flink 1.20

sigalit created FLINK-37228:
-------------------------------

             Summary: kubernetes HA fails both on flink 1.19.1 and flink 1.20 
                 Key: FLINK-37228
                 URL: https://issues.apache.org/jira/browse/FLINK-37228
             Project: Flink
          Issue Type: Bug
          Components: Runtime / Network
    Affects Versions: 1.19.1, 1.20.0, 1.19.0
            Reporter: sigalit
         Attachments: flinkdeployment.yaml, jobmanager1.log, jobmanager2.log, taskmanager.log

high-availability.type: kubernetes

state.checkpoint-storage: filesystem

state.checkpoints.dir: file://flink_checkpoints
state.savepoints.dir: file://flink_savepoints

the task manager is not able to connect to the kubernetes cluster to get the ha config map.

the job manager is able to create and watch the configmap, they are both using the service account and according to the logs the service account is taken.

when running the curl from the task manager pod the kubernetes api is working and gets the right data.

working with flink operator 1.9.

*on same kubernetes cluster when running with image of flink 1.18.1 no issue the HA is working fine.* 

attached logs and flinkdeployment yaml 

 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)