[
https://issues.apache.org/jira/browse/FLINK-20663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17276797#comment-17276797
]
Xintong Song commented on FLINK-20663:
--------------------------------------
I think option 2) might be a good way to go.
It looks like the underlying native memory of a segment is always accessed
through the wrapped ByteBuffers. We could implement something like a
WrappedByteBuffer. All read / write operations should be forwarded to the
original ByteBuffer before the underlying memory is released, and fail after
the memory is released. A ReadWriteLock is needed to make sure the freeing the
segment is visible to all the wrapped ByteBuffers immediately.
I'm still looking into this. There're a few things to be checked.
- Impact on performance due to the lock-based synchronization
- Potential read/write API changes on ByteBuffer between different Java versions
If this can work, I would suggest the following actions.
1. Fail explicitly on attempts to access the memory after it's freed
2. By default release the memory when the segment is freed. This would make the
potential access-after-release issues surface.
3. Keep the configuration option to disable releasing memory when the segment
is freed and fallback to GC, in case users indeed run into access-after-release
issues.
4. Remove the configuration option once we are confidence that there's no more
access-after-release issues. This is optional. We could also keep it incase new
access-after-release issues are introduced in future.
> Managed memory may not be released in time when operators use managed memory
> frequently
> ---------------------------------------------------------------------------------------
>
> Key: FLINK-20663
> URL: https://issues.apache.org/jira/browse/FLINK-20663
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Task
> Affects Versions: 1.12.0
> Reporter: Caizhi Weng
> Priority: Critical
> Fix For: 1.12.2
>
>
> Some batch operators (like sort merge join or hash aggregate) use managed
> memory frequently. When these operators are chained together and the cluster
> load is a bit heavy, it is very likely that the following exception occurs:
> {code:java}
> 2020-12-18 10:04:32
> java.lang.RuntimeException:
> org.apache.flink.runtime.memory.MemoryAllocationException: Could not allocate
> 512 pages
> at
> org.apache.flink.table.runtime.util.LazyMemorySegmentPool.nextSegment(LazyMemorySegmentPool.java:85)
> at
> org.apache.flink.runtime.io.disk.SimpleCollectingOutputView.<init>(SimpleCollectingOutputView.java:49)
> at
> org.apache.flink.table.runtime.operators.aggregate.BytesHashMap$RecordArea.<init>(BytesHashMap.java:297)
> at
> org.apache.flink.table.runtime.operators.aggregate.BytesHashMap.<init>(BytesHashMap.java:103)
> at
> org.apache.flink.table.runtime.operators.aggregate.BytesHashMap.<init>(BytesHashMap.java:90)
> at LocalHashAggregateWithKeys$209161.open(Unknown Source)
> at
> org.apache.flink.streaming.runtime.tasks.OperatorChain.initializeStateAndOpenOperators(OperatorChain.java:401)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.lambda$beforeInvoke$2(StreamTask.java:506)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTaskActionExecutor$SynchronizedStreamTaskActionExecutor.runThrowing(StreamTaskActionExecutor.java:92)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.beforeInvoke(StreamTask.java:501)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:530)
> at org.apache.flink.runtime.taskmanager.Task.doRun(Task.java:722)
> at org.apache.flink.runtime.taskmanager.Task.run(Task.java:547)
> at java.lang.Thread.run(Thread.java:834)
> Suppressed: java.lang.NullPointerException
> at LocalHashAggregateWithKeys$209161.close(Unknown Source)
> at
> org.apache.flink.table.runtime.operators.TableStreamOperator.dispose(TableStreamOperator.java:46)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.disposeAllOperators(StreamTask.java:739)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.runAndSuppressThrowable(StreamTask.java:719)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.cleanUpInvoke(StreamTask.java:642)
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask.invoke(StreamTask.java:551)
> ... 3 more
> Suppressed: java.lang.NullPointerException
> at LocalHashAggregateWithKeys$209766.close(Unknown
> Source)
> ... 8 more
> Caused by: org.apache.flink.runtime.memory.MemoryAllocationException: Could
> not allocate 512 pages
> at
> org.apache.flink.runtime.memory.MemoryManager.allocatePages(MemoryManager.java:231)
> at
> org.apache.flink.table.runtime.util.LazyMemorySegmentPool.nextSegment(LazyMemorySegmentPool.java:83)
> ... 13 more
> Caused by: org.apache.flink.runtime.memory.MemoryReservationException: Could
> not allocate 16777216 bytes, only 9961487 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:164)
> at
> org.apache.flink.runtime.memory.UnsafeMemoryBudget.reserveMemory(UnsafeMemoryBudget.java:80)
> at
> org.apache.flink.runtime.memory.MemoryManager.allocatePages(MemoryManager.java:229)
> ... 14 more
> {code}
> It seems that this is caused by relying on GC to release managed memory, as
> {{System.gc()}} may not trigger GC in time. See {{UnsafeMemoryBudget.java}}.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
