[
https://issues.apache.org/jira/browse/HBASE-19432?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16280113#comment-16280113
]
Hudson commented on HBASE-19432:
--------------------------------
FAILURE: Integrated in Jenkins build HBase-Trunk_matrix #4178 (See
[https://builds.apache.org/job/HBase-Trunk_matrix/4178/])
HBASE-19432 Roll the specified writer in HFileOutputFormat2 (tedyu: rev
6da52052ee37ac1ef74eca0030cc3951aad59f12)
* (edit)
hbase-mapreduce/src/main/java/org/apache/hadoop/hbase/mapreduce/HFileOutputFormat2.java
> Roll the specified writer in HFileOutputFormat2
> -----------------------------------------------
>
> Key: HBASE-19432
> URL: https://issues.apache.org/jira/browse/HBASE-19432
> Project: HBase
> Issue Type: Improvement
> Affects Versions: 3.0.0, 2.0.0-beta-1
> Reporter: Guangxu Cheng
> Assignee: Guangxu Cheng
> Fix For: 2.0.0-beta-1
>
> Attachments: HBASE-19432.master.001.patch,
> HBASE-19432.master.002.patch
>
>
> {code}
> // If any of the HFiles for the column families has reached
> // maxsize, we need to roll all the writers
> if (wl != null && wl.written + length >= maxsize) {
> this.rollRequested = true;
> }
> {code}
> If we always roll all the writers, a large number of small files will be
> generated in the multi family or multi table scene.
> So we should only roll the specified writer which HFile has reached maxsize.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
