[
https://issues.apache.org/jira/browse/YARN-4609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15217582#comment-15217582
]
Rohith Sharma K S commented on YARN-4609:
-----------------------------------------
Going through this ticket, Good observation [~bibinchundatt]:-) This
optimization is really helpful in large cluster.
The overall patch looks good to me.
One of the article I read about defferRedering after this issue is
[link|https://datatables.net/examples/ajax/defer_render.html]. They explain how
the tables are rendered fast if they enable defferRedering.
{code}
When deferred rendering is enabled, rather than having DataTables create all TR
and TD nodes required for the table when the data is loaded, DataTables will
only create the nodes required for each individual row at the time of that row
being drawn on the page (these nodes are then retained in case they are needed
again so they aren't created multiple times). This can give a significant
performance increase, since a lot less work is done at initialisation time.
{code}
I will simulate with and without patch, let me see the difference in page
loading time.
> RM Nodes list page takes too much time to load
> ----------------------------------------------
>
> Key: YARN-4609
> URL: https://issues.apache.org/jira/browse/YARN-4609
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Bibin A Chundatt
> Assignee: Bibin A Chundatt
> Attachments: 0001-YARN-4609.patch, 0002-YARN-4609.patch, 7k
> Nodes.png, sls-jobs.json, sls-nodes.json
>
>
> Configure SLS with 10000 NM Nodes
> Check the time taken to load Nodes page
> For loading 10 k Nodes it takes *30 sec*
> <RM IP>/cluster/nodes
> Chrome :Version 47.0.2526.106 m
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
