[
https://issues.apache.org/jira/browse/FLINK-36015?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rui Fan updated FLINK-36015:
----------------------------
Release Note:
FLIP-472 aligns timeout logic in AdaptiveScheduler states. To make alignment
more clear to users the configuration has also been alignment:
- Parameter `jobmanager.adaptive-scheduler.resource-wait-timeout` was renamed
to the `jobmanager.adaptive-scheduler.submission.resource-wait-timeout`
- Parameter `jobmanager.adaptive-scheduler.resource-stabilization-timeout` was
renamed to the
`jobmanager.adaptive-scheduler.submission.resource-stabilization-timeout`
- Parameter `jobmanager.adaptive-scheduler.scaling-interval.min` was renamed to
the `jobmanager.adaptive-scheduler.executing.cooldown-after-rescaling`
- Parameter `jobmanager.adaptive-scheduler.scaling-interval.max` was replaced
by the `jobmanager.adaptive-scheduler.executing.resource-stabilization-timeout`
with a default value of 60s.
- Parameter `jobmanager.adaptive-scheduler.min-parallelism-increase` was
removed without a replacement.
The following parameters introduced in Flink 2.0 have been also renamed:
- Parameter `jobmanager.adaptive-scheduler.max-delay-for-scale-trigger` was
renamed to the `jobmanager.adaptive-scheduler.rescale-trigger.max-delay`
- Parameter `jobmanager.adaptive-scheduler.scale-on-failed-checkpoints-count`
was renamed to the
`jobmanager.adaptive-scheduler.rescale-trigger.max-checkpoint-failures`
See Flink's configuration documentation for further details.
was:
FLIP-472 aligns timeout logic in AdaptiveScheduler states. To make alignment
more clear to users the configuration has also been alignment:
- Parameter `jobmanager.adaptive-scheduler.resource-wait-timeout` was renamed
to the `jobmanager.adaptive-scheduler.submission.resource-wait-timeout`
- Parameter `jobmanager.adaptive-scheduler.resource-stabilization-timeout` was
renamed to the
`jobmanager.adaptive-scheduler.submission.resource-stabilization-timeout`
- Parameter `jobmanager.adaptive-scheduler.scaling-interval.min` was renamed to
the `jobmanager.adaptive-scheduler.executing.cooldown-after-rescaling`
- Parameter `jobmanager.adaptive-scheduler.scaling-interval.max` was replaced
by the `jobmanager.adaptive-scheduler.executing.resource-stabilization-timeout`
with a default value of 60s.
- Parameter `jobmanager.adaptive-scheduler.min-parallelism-increase` was
removed without a replacement.
The following parameters introduced in Flink 2.0 have been also renamed:
- Parameter `jobmanager.adaptive-scheduler.max-delay-for-scale-trigger` was
renamed to the `jobmanager.adaptive-scheduler.rescale-trigger.max-delays`
- Parameter `jobmanager.adaptive-scheduler.scale-on-failed-checkpoints-count`
was renamed to the
`jobmanager.adaptive-scheduler.rescale-trigger.max-checkpoint-failures`
See Flink's configuration documentation for further details.
> Align rescale parameters
> ------------------------
>
> Key: FLINK-36015
> URL: https://issues.apache.org/jira/browse/FLINK-36015
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Configuration
> Reporter: Zdenek Tison
> Assignee: Zdenek Tison
> Priority: Major
> Labels: pull-request-available
> Fix For: 2.0-preview
>
>
> * Parameter
> [_jobmanager.adaptive-scheduler.resource-wait-timeout_|https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#jobmanager-adaptive-scheduler-resource-wait-timeout]
> will be renamed to the
> jobmanager.adaptive-scheduler.submission.resource-wait-timeout
> * Parameter
> [_jobmanager.adaptive-scheduler.resource-stabilization-timeout_|https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#jobmanager-adaptive-scheduler-resource-wait-timeout]
> will be renamed to the
> jobmanager.adaptive-scheduler.submission.resource-stabilization-timeout
> * Parameter
> {_}j{_}[_obmanager.adaptive-scheduler.scaling-interval.min_|https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#jobmanager-adaptive-scheduler-scaling-interval-min]
> will be renamed to the
> jobmanager.adaptive-scheduler.executing.cooldown-after-rescaling
> * Parameter
> [_jobmanager.adaptive-scheduler.scaling-interval.max_|https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#jobmanager-adaptive-scheduler-scaling-interval-max]
> will be renamed to the
> {_}jobmanager.adaptive-scheduler{_}{_}.{_}executing.resource-stabilization-timeout
> with default value 60s.
> * Parameter
> [jobmanager.adaptive-scheduler.min-parallelism-increase|https://nightlies.apache.org/flink/flink-docs-master/docs/deployment/config/#jobmanager-adaptive-scheduler-min-parallelism-increase]
> will be removed without a direct replacement. Still, it will be superseded
> by combining the parameters
> jobmanager.adaptive-scheduler.executing.cooldown-after-rescaling and
> {_}jobmanager.adaptive-scheduler{_}{_}.{_}executing.resource-stabilization-timeout
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
