[
https://issues.apache.org/jira/browse/HADOOP-2326?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sanjay Radia updated HADOOP-2326:
---------------------------------
Status: Patch Available (was: Open)
Submitting my last attached patch: dnBR_RandomBackoff5.txt
Tests on clusters of 450 nodes showed improvements on NN startup times.
If the NN's memory usage was high then the improvement in startup time can be
quite substantial.
Also retries of BRs was reduced.
> Add a Random backoff for the initial block report sent to the Name node
> -----------------------------------------------------------------------
>
> Key: HADOOP-2326
> URL: https://issues.apache.org/jira/browse/HADOOP-2326
> Project: Hadoop
> Issue Type: Sub-task
> Components: dfs
> Reporter: Sanjay Radia
> Assignee: Sanjay Radia
> Fix For: 0.16.0
>
> Attachments: dnBR_RandomBackoff2.txt, dnBR_RandomBackoff4.txt,
> dnBR_RandomBackoff5.txt
>
>
> Startup time can be improved if the initial block reports are spread randomly
> over small period of time.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
