hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sangjin Lee (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-6719) -libjars should use wildcards to reduce the application footprint in the state store
Date Tue, 21 Jun 2016 00:59:57 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-6719?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15340813#comment-15340813
] 

Sangjin Lee commented on MAPREDUCE-6719:
----------------------------------------

The latest patch looks good to me. The unit test that timed out is known to happen occasionally
(can't locate an open JIRA for that one though). The deprecation warning is also an unfortunately
result of {{DistributedCache}} being deprecated.

I do have a small quibble with the title of the JIRA. The title may give you an impression
that with this JIRA one could use a wildcard in specifying the libjars *argument*. But that's
not the case. What this JIRA does (in conjunction with YARN-4958) is to collapse the libjar
directory in the configuration as well as the container launch context. But the current title
can be confused with HADOOP-12747 ("support wildcard in libjars argument"). Can we modify
the title to better describe this JIRA?

> -libjars should use wildcards to reduce the application footprint in the state store
> ------------------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-6719
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-6719
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>          Components: distributed-cache
>    Affects Versions: 2.8.0
>            Reporter: Daniel Templeton
>            Assignee: Daniel Templeton
>            Priority: Critical
>         Attachments: MAPREDUCE-6719.001.patch, MAPREDUCE-6719.002.patch
>
>
> When using the -libjars option to add classes to the classpath, every library so added
is explicitly listed in the ContainerLaunchContext's local resources even though they're all
uploaded to the same directory in HDFS. When using tools like Crunch without an uber JAR or
when trying to take advantage of the shared cache, the number of libraries can be quite large.
We've seen many cases where we had to turn down the max number of applications to prevent
ZK from running out of heap because of the size of the state store entries.
> This JIRA proposes to allow for wildcards both in the internal processing of the -libjars
switch and in paths added through the Job and DistributedCache classes. Rather than listing
all files independently, this JIRA proposes to replace the complete list of libdir files with
the wildcarded libdir directory, e.g. "libdir/*". This behavior is the same as the current
behavior when using -libjars, but avoids explicitly listing every file.
> This capability will also be exposed by the {{DistributedCache.addCacheFile()}} method.
> See YARN-4958 for the NM side of the implementation and additional discussion.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-help@hadoop.apache.org


Mime
View raw message