You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by "Aihua Xu (Jira)" <ji...@apache.org> on 2020/09/16 03:09:00 UTC
[jira] [Created] (HIVE-24171) Support HDFS reads from observer
NameNodes
Aihua Xu created HIVE-24171:
-------------------------------
Summary: Support HDFS reads from observer NameNodes
Key: HIVE-24171
URL: https://issues.apache.org/jira/browse/HIVE-24171
Project: Hive
Issue Type: New Feature
Components: Hive
Affects Versions: 3.0.0
Reporter: Aihua Xu
Assignee: Aihua Xu
HDFS-12943 introduces the consistent reads from observer NameNodes which can boost the read performance and reduces the overloads on active NameNodes.
To take advantage of this feature, the clients are required to make a msync() call after writing the files or before reading the files since observer NameNodes could have the stale data for a small window.
Hive needs to make msync() call to HDFS in some places, e.g., 1) after generating the plan files - map.xml and reduce.xml so they can get used later by executors; 2) after the intermediate files are generated so they can get used by later stages or HS2.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)