[
https://issues.apache.org/jira/browse/FLINK-13223?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chesnay Schepler closed FLINK-13223.
------------------------------------
Resolution: Fixed
Fix Version/s: 1.10.0
Release Note:
Flink will now by default use the "region" failover strategy. Flink will now
attempt to only restart tasks that are connected to the failed task through a
pipelined connection.
Users who were not using a restart strategy or have already configured a
failover strategy should not be affected.
Streaming users who were not using a failover strategy may be affected if their
jobs are embarrassingly parallel or contain multiple independent jobs. In this
case, only the failed parallel pipeline or affected jobs will be restarted.
Batch users may be affected if their job contains blocking exchanges (usually
happens for shuffles) or the ExecutionMode was set to BATCH or BATCH_FORCED via
the ExecutionConfig.
Overall users should see an improvement in performance.
master: b610cedf18132ad11a44cc90a331349d3617c533
1.9: e436090f618ca4eadfd242c6b4fc1a97bbb35c16
> Set jobmanager.execution.failover-strategy to region in default
> flink-conf.yaml
> -------------------------------------------------------------------------------
>
> Key: FLINK-13223
> URL: https://issues.apache.org/jira/browse/FLINK-13223
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Coordination
> Reporter: Gary Yao
> Assignee: Chesnay Schepler
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 1.9.0, 1.10.0
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Set jobmanager.execution.failover-strategy to region in default
> flink-conf.yaml
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
