[
https://issues.apache.org/jira/browse/HIVE-21686?focusedWorklogId=241298&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-241298
]
ASF GitHub Bot logged work on HIVE-21686:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 13/May/19 20:46
Start Date: 13/May/19 20:46
Worklog Time Spent: 10m
Work Description: odraese commented on pull request #626: [HIVE-21686]
ensure that memory allocator does not evict using brute …
URL: https://github.com/apache/hive/pull/626#discussion_r283427024
##########
File path:
llap-server/src/java/org/apache/hadoop/hive/llap/cache/BuddyAllocator.java
##########
@@ -48,6 +48,24 @@
import org.apache.hadoop.hive.llap.metrics.LlapDaemonCacheMetrics;
import org.apache.hadoop.hive.ql.io.orc.encoded.StoppableAllocator;
+/**
+ *
+ * High level description, functionality and the memory layout.
+ * Allocation can be of general size but it will be rounded to the next power
of 2.
+ * Allocation smaller than size {@link ConfVars#LLAP_ALLOCATOR_MIN_ALLOC} will
be rounded to min allocation size.
+ * Allocation bigger than size {@link ConfVars#LLAP_ALLOCATOR_MAX_ALLOC} will
throw exceptions.
+ * Allocator slices memory slab called {@code arenas} to carve out byte
buffers using slice and position.
+ *
+ * Each {@code arena} has array of {@code freeList} used for concurrency
management and to index free buffers by size.
+ * |
+ * Each {@code arena} has a max Size {@link BuddyAllocator#MAX_ARENA_SIZE} 1GB.
+ * \_
+ * Each arena is divided into chunks of max allocation size {@link
ConfVars#LLAP_ALLOCATOR_MAX_ALLOC} default 16MB
+ * \_
+ * Each chunk of {@code maxAllocationSize} is sliced using the classical
Buddy Allocator algorithm.
+ * \_
+ * Each Buddy Allocator tree can be split up to chunk of size {@link
ConfVars#LLAP_ALLOCATOR_MIN_ALLOC} 4KB
Review comment:
chunk -> chunks
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 241298)
Time Spent: 2.5h (was: 2h 20m)
> Brute Force eviction can lead to a random uncontrolled eviction pattern.
> ------------------------------------------------------------------------
>
> Key: HIVE-21686
> URL: https://issues.apache.org/jira/browse/HIVE-21686
> Project: Hive
> Issue Type: Bug
> Reporter: slim bouguerra
> Assignee: slim bouguerra
> Priority: Major
> Labels: pull-request-available
> Attachments: Cache_hitrate_improvement.csv, HIVE-21686.2.patch,
> HIVE-21686.patch
>
> Time Spent: 2.5h
> Remaining Estimate: 0h
>
> Current logic used by brute force eviction can lead to a perpetual random
> eviction pattern.
> For instance if the cache build a small pocket of free memory where the total
> size is greater than incoming allocation request, the allocator will randomly
> evict block that fits a particular size.
> This can happen over and over therefore all the eviction will be random.
> In Addition this random eviction will lead a leak in the linked list
> maintained by the policy since it does not know anymore about what is evicted
> and what not.
> The improvement of this patch is very substantial to TPC-DS benchmark. I
> have tested it with 10TB scale 9 llap nodes and 32GB cache size per node.
> The patch has showed very noticeable difference in the Hit rate for raw
> number [^Cache_hitrate_improvement.csv]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
