[
https://issues.apache.org/jira/browse/FLINK-7499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16264554#comment-16264554
]
ASF GitHub Bot commented on FLINK-7499:
---------------------------------------
Github user pnowojski commented on a diff in the pull request:
https://github.com/apache/flink/pull/4581#discussion_r152825430
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/io/network/partition/ResultPartition.java
---
@@ -267,28 +268,29 @@ public ResultPartitionType getPartitionType() {
* first buffer has been added.
*/
public void add(Buffer buffer, int subpartitionIndex) throws
IOException {
- boolean success = false;
+ checkNotNull(buffer);
try {
checkInProduceState();
+ } catch (Throwable t) {
+ buffer.recycle();
--- End diff --
I wonder if we should have some sanity illegal state detection for double
recycling the buffers. For example each buffer could only be recycled once
(protected by a private field in the Buffer `boolean wasRecycled`). Whenever
you call `retain()`, you would get a new instance of the `Buffer`, pointing to
the same memory, but with new flag (so that both original and retained buffers
could be recycled independently, but each one of them only once).
> double buffer release in SpillableSubpartitionView
> --------------------------------------------------
>
> Key: FLINK-7499
> URL: https://issues.apache.org/jira/browse/FLINK-7499
> Project: Flink
> Issue Type: Sub-task
> Components: Network
> Affects Versions: 1.2.0, 1.1.4, 1.3.0, 1.1.5, 1.2.1, 1.2.2, 1.3.1, 1.4.0,
> 1.3.2, 1.3.3
> Reporter: Nico Kruber
> Assignee: Nico Kruber
> Priority: Blocker
>
> {{SpillableSubpartitionView#releaseMemory()}} recycles its buffers twice:
> once asynchronously after the write operation and once in
> {{SpillableSubpartitionView#releaseMemory()}} after adding the write
> operation to the queue.
> 1) if {{SpillableSubpartitionView#releaseMemory()}} hits first and the buffer
> is recycled, the memory region may already be reused despite the pending write
> 2) If, for some reason (probably only in tests like
> {{SpillableSubpartitionTest#testConsumeSpillablePartitionSpilledDuringConsume()}}?),
> the buffer is retained and to be used in parallel somewhere else it may also
> not be available anymore or contain corrupt data.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
