timarmstrong commented on a change in pull request #34186:
URL: https://github.com/apache/spark/pull/34186#discussion_r731153656
##########
File path: core/src/main/java/org/apache/spark/memory/TaskMemoryManager.java
##########
@@ -135,109 +135,117 @@ public TaskMemoryManager(MemoryManager memoryManager,
long taskAttemptId) {
*
* @return number of bytes successfully granted (<= N).
*/
- public long acquireExecutionMemory(long required, MemoryConsumer consumer) {
+ public long acquireExecutionMemory(long required, MemoryConsumer
requestingConsumer) {
assert(required >= 0);
- assert(consumer != null);
- MemoryMode mode = consumer.getMode();
+ assert(requestingConsumer != null);
+ MemoryMode mode = requestingConsumer.getMode();
// If we are allocating Tungsten pages off-heap and receive a request to
allocate on-heap
// memory here, then it may not make sense to spill since that would only
end up freeing
// off-heap memory. This is subject to change, though, so it may be risky
to make this
// optimization now in case we forget to undo it late when making changes.
synchronized (this) {
+ consumers.add(requestingConsumer);
long got = memoryManager.acquireExecutionMemory(required, taskAttemptId,
mode);
Review comment:
The `c.getUsed() > 0` check also should prevent the requesting consumer
being added to the map on the first call.
It sounds like there's a consensus that we add it at the end, like the
original code? I'll revert my change then.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
