JoshRosen commented on a change in pull request #25953: [SPARK-29244][Core]
Prevent freed page in BytesToBytesMap free again
URL: https://github.com/apache/spark/pull/25953#discussion_r329324260
##########
File path: core/src/main/java/org/apache/spark/unsafe/map/BytesToBytesMap.java
##########
@@ -787,8 +783,16 @@ private void allocate(int capacity) {
assert (capacity >= 0);
capacity = Math.max((int) Math.min(MAX_CAPACITY,
ByteArrayMethods.nextPowerOf2(capacity)), 64);
assert (capacity <= MAX_CAPACITY);
- longArray = allocateArray(capacity * 2L);
- longArray.zeroOut();
+ try {
+ longArray = allocateArray(capacity * 2L);
+ longArray.zeroOut();
+ } catch (SparkOutOfMemoryError e) {
+ // When OOM, allocated page was already freed by `TaskMemoryManager`.
+ // We should not keep it in `longArray`. Otherwise it might be freed
again in task
+ // complete listeners and cause unnecessary error.
+ longArray = null;
Review comment:
In retrospect, with a few years of hindsight, I think there's definitely
room to write `BytesToBytesMap` in an easier-to-understand manner: it's quite
possible that a different design of helper methods and management of
side-effects would prevent this class of bugs or make them easier to fix. For
the scope of this patch, though, I think we should pursue a surgical fix (with
re-architecture considered later).
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
