[SPARK-22827][CORE] Avoid throwing OutOfMemoryError in case of exception in spill #20014
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Jenkins test this please. |
|
cc - @rxin, @sameeragarwal, @zsxwing, |
|
Test build #85080 has finished for PR 20014 at commit
|
|
Test build #85081 has finished for PR 20014 at commit
|
rxin
reviewed
Dec 19, 2017
| * Instead of throwing {@link OutOfMemoryError}, which kills the executor, | ||
| * we should use throw this exception, which will just kill the current task. | ||
| */ | ||
| public final class SparkOutOfMemoryError extends OutOfMemoryError { |
There was a problem hiding this comment.
is this an internal class? if yes perhaps we should label it.
There was a problem hiding this comment.
Yes, this is an internal class. How do you suggest to label it ?
There was a problem hiding this comment.
import org.apache.spark.annotation.Private;
@Private
|
Overall change lgtm. |
|
Test build #85102 has finished for PR 20014 at commit
|
jiangxb1987
approved these changes
Dec 19, 2017
| /** | ||
| * This exception is thrown when a task can not acquire memory from the Memory manager. | ||
| * Instead of throwing {@link OutOfMemoryError}, which kills the executor, | ||
| * we should use throw this exception, which will just kill the current task. |
There was a problem hiding this comment.
nit: we should throw this exception, which just kills the current task.
cloud-fan
reviewed
Dec 19, 2017
| import org.apache.spark.unsafe.array.LongArray; | ||
| import org.apache.spark.unsafe.memory.MemoryBlock; | ||
|
|
||
| import java.io.IOException; |
There was a problem hiding this comment.
java import should comes first, please keep the previous code.
cloud-fan
reviewed
Dec 19, 2017
| @@ -341,7 +342,7 @@ private void growPointerArrayIfNecessary() throws IOException { | |||
| // should have trigger spilling | |||
| if (!inMemSorter.hasSpaceForAnotherRecord()) { | |||
| logger.error("Unable to grow the pointer array"); | |||
| throw e; | |||
| throw new SparkOutOfMemoryError(e); | |||
There was a problem hiding this comment.
Should allocateArray also throw SparkOutOfMemoryError then we can just rethrow it here?
There was a problem hiding this comment.
Actually, it already throws SparkOutOfMemoryError, so we can just rethrow here.
|
LGTM except 2 minor comments |
added 2 commits
December 19, 2017 10:42
|
Test build #85115 has finished for PR 20014 at commit
|
|
Test build #85127 has finished for PR 20014 at commit
|
|
thanks, merging to master! |
asfgit
pushed a commit
that referenced
this pull request
Nov 8, 2018
…ggregateExec`, too. ## What changes were proposed in this pull request? This is a follow-up pr of #20014 which introduced `SparkOutOfMemoryError` to avoid killing the entire executor when an `OutOfMemoryError` is thrown. We should throw `SparkOutOfMemoryError` in `HashAggregateExec`, too. ## How was this patch tested? Existing tests. Closes #22969 from ueshin/issues/SPARK-22827/oome. Authored-by: Takuya UESHIN <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
asfgit
pushed a commit
that referenced
this pull request
Nov 23, 2018
…moryError when catch exception ## What changes were proposed in this pull request? the pr #20014 which introduced `SparkOutOfMemoryError` to avoid killing the entire executor when an `OutOfMemoryError `is thrown. so apply for memory using `MemoryConsumer. allocatePage `when catch exception, use `SparkOutOfMemoryError `instead of `OutOfMemoryError` ## How was this patch tested? N / A Closes #23084 from heary-cao/SparkOutOfMemoryError. Authored-by: caoxuewen <[email protected]> Signed-off-by: Wenchen Fan <[email protected]>
jackylee-ch
pushed a commit
to jackylee-ch/spark
that referenced
this pull request
Feb 18, 2019
…ggregateExec`, too. ## What changes were proposed in this pull request? This is a follow-up pr of apache#20014 which introduced `SparkOutOfMemoryError` to avoid killing the entire executor when an `OutOfMemoryError` is thrown. We should throw `SparkOutOfMemoryError` in `HashAggregateExec`, too. ## How was this patch tested? Existing tests. Closes apache#22969 from ueshin/issues/SPARK-22827/oome. Authored-by: Takuya UESHIN <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
jackylee-ch
pushed a commit
to jackylee-ch/spark
that referenced
this pull request
Feb 18, 2019
…moryError when catch exception ## What changes were proposed in this pull request? the pr apache#20014 which introduced `SparkOutOfMemoryError` to avoid killing the entire executor when an `OutOfMemoryError `is thrown. so apply for memory using `MemoryConsumer. allocatePage `when catch exception, use `SparkOutOfMemoryError `instead of `OutOfMemoryError` ## How was this patch tested? N / A Closes apache#23084 from heary-cao/SparkOutOfMemoryError. Authored-by: caoxuewen <[email protected]> Signed-off-by: Wenchen Fan <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Currently, the task memory manager throws an OutofMemory error when there is an IO exception happens in spill() - https://github.com/apache/spark/blob/master/core/src/main/java/org/apache/spark/memory/TaskMemoryManager.java#L194. Similarly there any many other places in code when if a task is not able to acquire memory due to an exception we throw an OutofMemory error which kills the entire executor and hence failing all the tasks that are running on that executor instead of just failing one single task.
How was this patch tested?
Unit tests