Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/16603#discussion_r96565501
--- Diff: core/src/main/java/org/apache/spark/memory/TaskMemoryManager.java
---
@@ -144,23 +170,31 @@ public long acquireExecutionMemory(long required,
MemoryConsumer consumer) {
// spilling, avoid to have too many spilled files.
if (got < required) {
// Call spill() on other consumers to release memory
+ // Sort the consumers according their memory usage. So we avoid
spilling the same consumer
+ // which is just spilled in last few times and re-spilling on it
will produce many small
+ // spill files.
+ List<MemoryConsumer> sortedList = new ArrayList<>();
for (MemoryConsumer c: consumers) {
if (c != consumer && c.getUsed() > 0 && c.getMode() == mode) {
- try {
- long released = c.spill(required - got, consumer);
- if (released > 0) {
- logger.debug("Task {} released {} from {} for {}",
taskAttemptId,
- Utils.bytesToString(released), c, consumer);
- got += memoryManager.acquireExecutionMemory(required -
got, taskAttemptId, mode);
- if (got >= required) {
- break;
- }
+ sortedList.add(c);
+ }
+ }
+ Collections.sort(sortedList, new ConsumerComparator());
+ for (MemoryConsumer c: sortedList) {
+ try {
+ long released = c.spill(required - got, consumer);
+ if (released > 0) {
+ logger.debug("Task {} released {} from {} for {}",
taskAttemptId,
+ Utils.bytesToString(released), c, consumer);
+ got += memoryManager.acquireExecutionMemory(required - got,
taskAttemptId, mode);
+ if (got >= required) {
+ break;
}
- } catch (IOException e) {
- logger.error("error while calling spill() on " + c, e);
- throw new OutOfMemoryError("error while calling spill() on "
+ c + " : "
- + e.getMessage());
}
+ } catch (IOException e) {
+ logger.error("error while calling spill() on " + c, e);
+ throw new OutOfMemoryError("error while calling spill() on " +
c + " : "
+ + e.getMessage());
}
--- End diff --
I don't think the best or better strategy is to spill many smaller
consumers for a bigger memory request.
Too frequently spill small consumers should cause GC pressure as reported
in #16387. Actually this PR is inspired by that to avoid the same consumer is
spilled again and again even it only uses small memory.
The newest update is to spill the ceiling entry (the least consumer uses
greater than or equal to the required memory).
For the same example, consumers = 50 MB, 100MB, 200 MB, 250MB, 500 MB, 1GB,
it behaves like:
Required: 1.4 GB
Evict: 1GB, 500MB
Required: 300 MB
Evict: 500MB
Required: 400MB
Evict: 500MB
Required: 60 MB
Evict 100 MB
Required: 200 MB
Evict 200 MB
It is nearly the same as the result you showed. As I pointed in previous
comment, the logic you showed actually can't produce that result.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
