[
https://issues.apache.org/jira/browse/HDFS-14635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17953241#comment-17953241
]
Dheeren Beborrtha commented on HDFS-14635:
------------------------------------------
Why is this not part of regular release ? This is an important bug/feature. We
should avoid restart of namenode.
In clusters where new hdfs datanodes are to be added, this needs a restart of
namenode. And this could be an expensive operation in a large loaded clusters.
> Support to refresh the rack awareness dynamically
> --------------------------------------------------
>
> Key: HDFS-14635
> URL: https://issues.apache.org/jira/browse/HDFS-14635
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs
> Affects Versions: 2.7.2
> Reporter: liying
> Priority: Major
>
> At present , there are two ways to load the rack script in the hadoop codes.
> The class ScriptBasedMapping is the cache way, and the class
> ScriptBasedMapping#RawScriptBasedMapping is the way that
> it will load script every time（every request）。 The cache is the good way to
> implement this feature， because it consumes cpu performance if loading the
> script for every quest。But here's another question that is we cann't refresh
> the cache, so it is import to support to refresh the rack awareness
> dynamically.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
