[
https://issues.apache.org/jira/browse/DRILL-3811?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14901697#comment-14901697
]
ASF GitHub Bot commented on DRILL-3811:
---------------------------------------
Github user amansinha100 commented on a diff in the pull request:
https://github.com/apache/drill/pull/163#discussion_r40045182
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/memory/AtomicRemainder.java
---
@@ -76,16 +74,17 @@ public void setLimit(long limit) {
}
/**
- * Automatically allocate memory. This is used when an actual allocation
happened to be larger than requested. This
- * memory has already been used up so it must be accurately accounted
for in future allocations.
+ * Automatically allocate memory. This is used when an actual allocation
happened to be larger than requested, or when
+ * a buffer has it's ownership passed to another allocator.<br>
+ * This memory has already been used up so it must be accurately
accounted for in future allocations.
*
- * @param size
+ * @param size extra allocated memory that needs to be accounted for
*/
public boolean forceGet(long size) {
if (get(size, this.applyFragmentLimit)) {
return true;
} else {
- availableShared.addAndGet(size);
+ availableShared.addAndGet(-size);
--- End diff --
Since this will be subtracting size bytes, should there be a check for
availableShared >=0 ? I am not quite sure what's supposed to happen if this
value drops below 0 at this point...
> AtomicRemainder incorrectly accounts for transferred allocations
> ----------------------------------------------------------------
>
> Key: DRILL-3811
> URL: https://issues.apache.org/jira/browse/DRILL-3811
> Project: Apache Drill
> Issue Type: Bug
> Components: Execution - Relational Operators
> Reporter: Deneche A. Hakim
> Assignee: Chris Westin
> Attachments: DRILL-3811.1.patch.txt
>
>
> when an allocator takes ownership of a buffer, AtomicRemainder.forceGet(int)
> is called to account for the extra memory of the buffer, but when the
> allocator exceeds it's maximum allocated memory it accounts for it
> incorrectly. In the following code, {{availableShared.andAndGet(size)}}
> should actually receive {{-size}}:
> {code}
> public boolean forceGet(long size) {
> if (get(size, this.applyFragmentLimit)) {
> return true;
> } else {
> availableShared.addAndGet(size);
> if (parent != null) {
> parent.forceGet(size);
> }
> return false;
> }
> }
> {code}
> I was able to reproduce the issue in a simple unit test
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
