[SPARK-10582] using dynamic-executor-allocation, if AM failed. the new AM will be started. But the new AM does not allocate executors to dirver

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.4.1, 1.5.1
• Fix Version/s: 2.0.0
• Component/s: Spark Core
• Assignee: Saisai Shao
• Reporter: KaiXinXIaoLei
• Due: 14/Sep/15
• Created: 14/Sep/15 02:42
• Updated: 23/Feb/16 07:17
• Resolved: 09/Dec/15 17:52

Description

During running tasks, when the total number of executors is the value of spark.dynamicAllocation.maxExecutors and the AM is failed. Then a new AM restarts. Because in ExecutorAllocationManager, the total number of executors does not changed, driver does not send RequestExecutors to AM to ask executors. Then the total number of executors is the value of spark.dynamicAllocation.initialExecutors . So the total number of executors in driver and AM is different.

Issue Links

• is related to: SPARK-13447 Fix AM failure situation for dynamic allocation disabled situation
• links to: [Github] Pull Request #8737 (KaiXinXiaoLei)
• links to: [Github] Pull Request #9963 (jerryshao)