Nuno Afonso created FLINK-28651:
-----------------------------------
Summary: JarRunHandler ignores restore mode set in the
configuration
Key: FLINK-28651
URL: https://issues.apache.org/jira/browse/FLINK-28651
Project: Flink
Issue Type: Bug
Components: Runtime / REST
Affects Versions: 1.15.1, 1.15.0
Reporter: Nuno Afonso
Hello all,
I started a Flink cluster with execution.savepoint-restore-mode set to CLAIM.
Then, I submitted a job through the REST API without specifying restoreMode in
the request. I would expect Flink to use CLAIM, but the job used NO_CLAIM as
restore mode.
After looking into the source code, I believe the issue comes from the way
JarRunHandler picks the default for restoreMode. It directly gets it from the
default of execution.savepoint-restore-mode instead of looking into the
existing configuration:
[https://github.com/apache/flink/blob/release-1.15.1/flink-runtime-web/src/main/java/org/apache/flink/runtime/webmonitor/handlers/JarRunHandler.java#L150]
I think the fix can be achieved by getting execution.savepoint-restore-mode
from the configuration.
Looking forward to hearing your feedback.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
