[FLINK-7805] Add HA capabilities to YarnResourceManager

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.4.0
• Fix Version/s: 1.5.0
• Component/s: Deployment / YARN, Runtime / Coordination
• Labels: flip-6
• Assignee: Gary Yao
• Reporter: Till Rohrmann
• Created: 11/Oct/17 08:41
• Updated: 17/May/18 13:31
• Resolved: 15/Mar/18 16:56

Description

The new YarnResourceManager implementation does not retrieve allocated containers from previous attempts in HA mode like the old YarnFlinkResourceManager did. We should add this functionality in order to properly support long running Yarn applications [1].

[1] https://de.hortonworks.com/blog/apache-hadoop-yarn-hdp-2-2-fault-tolerance-features-long-running-services/

Issue Links

• is duplicated by: FLINK-8434 The new yarn resource manager should take over the running task managers after failover
• links to: GitHub Pull Request #5597