[FLINK-21552] The managed memory was not released if exception was thrown in createPythonExecutionEnvironment

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.12.0
• Fix Version/s: 1.13.0, 1.12.3
• Component/s: API / Python, Runtime / Coordination
• Labels: pull-request-available
• Assignee: Xintong Song
• Reporter: Dian Fu
• Created: 02/Mar/21 06:39
• Updated: 04/Mar/21 06:04
• Resolved: 04/Mar/21 06:04

Description

If there is exception thrown in createPythonExecutionEnvironment, the job will failed with the following exception:

org.apache.flink.runtime.memory.MemoryAllocationException: Could not created the shared memory resource of size 611948962. Not enough memory left to reserve from the slot's managed memory.
at org.apache.flink.runtime.memory.MemoryManager.lambda$getSharedMemoryResourceForManagedMemory$5(MemoryManager.java:536)
at org.apache.flink.runtime.memory.SharedResources.createResource(SharedResources.java:126)
at org.apache.flink.runtime.memory.SharedResources.getOrAllocateSharedResource(SharedResources.java:72)
at org.apache.flink.runtime.memory.MemoryManager.getSharedMemoryResourceForManagedMemory(MemoryManager.java:555)
at org.apache.flink.streaming.api.runners.python.beam.BeamPythonFunctionRunner.open(BeamPythonFunctionRunner.java:250)
at org.apache.flink.streaming.api.operators.python.AbstractPythonFunctionOperator.open(AbstractPythonFunctionOperator.java:113)
at org.apache.flink.table.runtime.operators.python.AbstractStatelessFunctionOperator.open(AbstractStatelessFunctionOperator.java:116)
at org.apache.flink.table.runtime.operators.python.scalar.AbstractPythonScalarFunctionOperator.open(AbstractPythonScalarFunctionOperator.java:88)
at org.apache.flink.table.runtime.operators.python.scalar.AbstractRowDataPythonScalarFunctionOperator.open(AbstractRowDataPythonScalarFunctionOperator.java:70)
at org.apache.flink.table.runtime.operators.python.scalar.RowDataPythonScalarFunctionOperator.open(RowDataPythonScalarFunctionOperator.java:59)
at org.apache.flink.streaming.runtime.tasks.OperatorChain.initializeStateAndOpenOperators(OperatorChain.java:428)
at org.apache.flink.streaming.runtime.tasks.StreamTask.lambda$beforeInvoke$2(StreamTask.java:543)
at org.apache.flink.streaming.runtime.tasks.StreamTaskActionExecutor$SynchronizedStreamTaskActionExecutor.runThrowing(StreamTaskActionExecutor.java:93)
at org.apache.flink.streaming.runtime.tasks.StreamTask.beforeInvoke(StreamTask.java:533)
at org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:573)
at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:755)
at org.apache.flink.runtime.taskmanager.Task.run(Task.java:570)
at java.lang.Thread.run(Thread.java:834)
Caused by: org.apache.flink.runtime.memory.MemoryReservationException: Could not allocate 611948962 bytes, only 0 bytes are remaining. This usually indicates that you are requesting more memory than you have reserved. However, when running an old JVM version it can also be caused by slow garbage collection. Try to upgrade to Java 8u72 or higher if running on an old Java version.
at org.apache.flink.runtime.memory.UnsafeMemoryBudget.reserveMemory(UnsafeMemoryBudget.java:170)
at org.apache.flink.runtime.memory.UnsafeMemoryBudget.reserveMemory(UnsafeMemoryBudget.java:84)
at org.apache.flink.runtime.memory.MemoryManager.reserveMemory(MemoryManager.java:423)
at org.apache.flink.runtime.memory.MemoryManager.lambda$getSharedMemoryResourceForManagedMemory$5(MemoryManager.java:534)
... 17 more

The reason is that the reserved managed memory was not added back to the MemoryManager when Job failed because of exceptions thrown in createPythonExecutionEnvironment. This causes that there is no managed memory to allocate during failover.

Issue Links

• links to: GitHub Pull Request #15057