[
https://issues.apache.org/jira/browse/HBASE-13160?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14350757#comment-14350757
]
Jeffrey Zhong commented on HBASE-13160:
---------------------------------------
+1. Two very minor things:
if we could append the condition
{code}
if (seq_start == taskReadySeq) {
{code}
to
{code}
if (seq_start == taskReadySeq && numTasks==0) {
{code}
2) the following isn't needed any more
{noformat}
if (childrenPaths != null) {
return childrenPaths;
}
{noformat}
> SplitLogWorker does not pick up the task immediately
> ----------------------------------------------------
>
> Key: HBASE-13160
> URL: https://issues.apache.org/jira/browse/HBASE-13160
> Project: HBase
> Issue Type: Bug
> Reporter: Enis Soztutar
> Assignee: Enis Soztutar
> Fix For: 2.0.0, 1.1.0
>
> Attachments: hbase-13160_v1.patch
>
>
> We were reading some code with Jeffrey, and we realized that the
> SplitLogWorker's internal task loop is weird. It does {{ls}} every second and
> sleeps, but have another mechanism to learn about new tasks, but does not
> make affective use of the zk notification.
> I have a simple patch which might improve this area.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)