[
https://issues.apache.org/jira/browse/FLINK-14894?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16982499#comment-16982499
]
Andrey Zagrebin commented on FLINK-14894:
-----------------------------------------
I think this is due to
[FLINK-13985|https://jira.apache.org/jira/browse/FLINK-13985], where we decided
that we will explicitly release underlying unsafe memory in
HybridMemorySegment#free for new unsafe allocations. It looks like
MemorySegmentTestBase#testByteBufferWrapping explicitly tests the contract that
the underlying memory wrapped into a nio ByteBuffer is usable after the segment
freeing. This is true for the nio DirectByteBuffer as the wrapping ByteBuffer
keeps a strong link to it. So the original memory is not released both
explicitly and during GC in this case but this does not hold for the new unsafe
allocation after the explicit release.
[~sewen] do you remember why would we need the wraps to be valid after freeing
the main segment? do we still need this?
> HybridOffHeapUnsafeMemorySegmentTest#testByteBufferWrap failed on Travis
> ------------------------------------------------------------------------
>
> Key: FLINK-14894
> URL: https://issues.apache.org/jira/browse/FLINK-14894
> Project: Flink
> Issue Type: Bug
> Components: Tests
> Affects Versions: 1.10.0
> Reporter: Gary Yao
> Priority: Critical
>
> {noformat}
> HybridOffHeapUnsafeMemorySegmentTest>MemorySegmentTestBase.testByteBufferWrapping:2465
> expected:<992288337> but was:<196608>
> {noformat}
> https://api.travis-ci.com/v3/job/258950527/log.txt
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
