[
https://issues.apache.org/jira/browse/PIG-4299?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Prashant Kommireddi updated PIG-4299:
-------------------------------------
Description:
{code}
for (MemoryPoolMXBean b: mpbeans) {
log.debug("Found heap (" + b.getName() +
") of type " + b.getType());
if (b.getType() == MemoryType.HEAP) {
/* Here we are making the leap of faith that the biggest
* heap is the tenured heap
*/
long size = b.getUsage().getMax();
totalSize += size;
if (size > biggestSize) {
biggestSize = size;
biggestHeap = b;
}
}
}
{code}
A memory pool being the biggest MemoryType.HEAP does not guarantee it being
tenured? Moreover, we must check whether usage threshold is supported by heap
before trying to set usage threshold on it.
was:
{code}
for (MemoryPoolMXBean b: mpbeans) {
log.debug("Found heap (" + b.getName() +
") of type " + b.getType());
if (b.getType() == MemoryType.HEAP) {
/* Here we are making the leap of faith that the biggest
* heap is the tenured heap
*/
long size = b.getUsage().getMax();
totalSize += size;
if (size > biggestSize) {
biggestSize = size;
biggestHeap = b;
}
}
}
{code}
A memory pool being MemoryType.HEAP does not guarantee it being tenured.
Moreover, we must check whether usage threshold is supported by heap before
trying to set usage threshold on it.
> SpillableMemoryManager assumes tenured heap incorrectly
> -------------------------------------------------------
>
> Key: PIG-4299
> URL: https://issues.apache.org/jira/browse/PIG-4299
> Project: Pig
> Issue Type: Bug
> Affects Versions: 0.13.0
> Reporter: Prashant Kommireddi
> Fix For: 0.14.0
>
>
> {code}
> for (MemoryPoolMXBean b: mpbeans) {
> log.debug("Found heap (" + b.getName() +
> ") of type " + b.getType());
> if (b.getType() == MemoryType.HEAP) {
> /* Here we are making the leap of faith that the biggest
> * heap is the tenured heap
> */
> long size = b.getUsage().getMax();
> totalSize += size;
> if (size > biggestSize) {
> biggestSize = size;
> biggestHeap = b;
> }
> }
> }
> {code}
> A memory pool being the biggest MemoryType.HEAP does not guarantee it being
> tenured? Moreover, we must check whether usage threshold is supported by heap
> before trying to set usage threshold on it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
