[
https://issues.apache.org/jira/browse/FLINK-31406?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gyula Fora closed FLINK-31406.
------------------------------
Resolution: Won't Do
The gain from this feature is pretty limited as newer Flink versions all
support in-place scaling, at the same time the introduced complexity in the
operator flow can be quite significant. We decided to not do this for now
> Do not delete jobgraph on scale only last-state upgrades
> --------------------------------------------------------
>
> Key: FLINK-31406
> URL: https://issues.apache.org/jira/browse/FLINK-31406
> Project: Flink
> Issue Type: Improvement
> Components: Kubernetes Operator
> Reporter: Gyula Fora
> Priority: Major
> Labels: pull-request-available, stale-assigned
>
> Currently the operator always deletes the jobgraph from HA metadata so that
> it's regenerated for last-state upgrades.
> This is unnecessary for scale only operations. Keeping the jobgraph can
> greately speed up startup time for some jobs.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
