[
https://issues.apache.org/jira/browse/HIVE-16180?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Prasanth Jayachandran updated HIVE-16180:
-----------------------------------------
Attachment: Native-mem-spike.png
FullGC-15GB-cleanup.png
Full-gc-native-mem-cleanup.png
Attaching snapshots from the analysis.
1) Native memory spike happened at around 11:00:00PM (configured offheap cache
size is 48GB). At this point memory used by direct byte buffer spiked to
51.7GB. !Native-mem-spike.png!
2) At around same time Full GC triggered and reclaimed 15GB of memory (12GB
heap + 3 GB offheap) !FullGC-15GB-cleanup.png! !Full-gc-native-mem-cleanup.png!
> LLAP: Native memory leak in EncodedReader
> -----------------------------------------
>
> Key: HIVE-16180
> URL: https://issues.apache.org/jira/browse/HIVE-16180
> Project: Hive
> Issue Type: Bug
> Components: llap
> Affects Versions: 2.2.0
> Reporter: Prasanth Jayachandran
> Assignee: Prasanth Jayachandran
> Priority: Critical
> Attachments: DirectCleaner.java, FullGC-15GB-cleanup.png,
> Full-gc-native-mem-cleanup.png, HIVE-16180.1.patch, Native-mem-spike.png
>
>
> Observed this in internal test run. There is a native memory leak in Orc
> EncodedReaderImpl that can cause YARN pmem monitor to kill the container
> running the daemon. Direct byte buffers are null'ed out which is not
> guaranteed to be cleaned until next Full GC. To show this issue, attaching a
> small test program that allocates 3x256MB direct byte buffers. First buffer
> is null'ed out but still native memory is used. Second buffer user Cleaner to
> clean up native allocation. Third buffer is also null'ed but this time
> invoking a System.gc() which cleans up all native memory. Output from the
> test program is below
> {code}
> Allocating 3x256MB direct memory..
> Native memory used: 786432000
> Native memory used after data1=null: 786432000
> Native memory used after data2.clean(): 524288000
> Native memory used after data3=null: 524288000
> Native memory used without gc: 524288000
> Native memory used after gc: 0
> {code}
> Longer term improvements/solutions:
> 1) Use DirectBufferPool from hadoop or netty's
> https://netty.io/4.0/api/io/netty/buffer/PooledByteBufAllocator.html as
> direct byte buffer allocations are expensive (System.gc() + 100ms thread
> sleep).
> 2) Use HADOOP-12760 for proper cleaner invocation in JDK8 and JDK9
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
