You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@tez.apache.org by "Syed Shameerur Rahman (Jira)" <ji...@apache.org> on 2023/01/31 05:27:00 UTC
[jira] [Commented] (TEZ-4397) Open Tez Input splits asynchronously
[ https://issues.apache.org/jira/browse/TEZ-4397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17682391#comment-17682391 ]
Syed Shameerur Rahman commented on TEZ-4397:
--------------------------------------------
[~abstractdog]
On debugging i found that This was due Hive’s operator initialization which are designed in a way that only one split can be opened at a time since they share the [IOContext|https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/io/IOContext.java] variable so when multiple splits are opened at the same time the shared variable gets overwritten and doesn’t match the correct context.
I will take this up separately in Hive, But for Tez I have created a new PR with the following changes
1. Disable the feature by default (no. of reader = 1)
> Open Tez Input splits asynchronously
> ------------------------------------
>
> Key: TEZ-4397
> URL: https://issues.apache.org/jira/browse/TEZ-4397
> Project: Apache Tez
> Issue Type: Task
> Reporter: Ramesh Kumar Thangarajan
> Assignee: Ramesh Kumar Thangarajan
> Priority: Major
> Time Spent: 2h 40m
> Remaining Estimate: 0h
>
> Tez input splits can be opened asynchronously. This will reduce the amount of time spent for s3 to prepare the connection and opening the object.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)