[
https://issues.apache.org/jira/browse/HADOOP-2916?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12573565#action_12573565
]
Arun C Murthy commented on HADOOP-2916:
---------------------------------------
{quote}
src/core - this will contain the core classes that hadoop applications need to
link against.
It will contain client side libraries of all fs file systems: local, hdfs, kfs,
etc
jar name hadoop_core.jar
{quote}
I really don't think the local, hdfs, kfs clients belong in *core* at all. It
doesn't make sense to have Map-Reduce depend on the kfs client libraries...
We really should be having {hdfs|kfs}_server.jar and
{hdfs|kfs|local}_client.jar. Thoughts?
Maybe I can be persuaded to let local-fs_client in hadoop-core.jar! *smile*
The one other view is to have a fs_core.jar which contains only the fs
interface and other file-system generics. This way hadoop-core is really just
the 'core' infrastructure.. and projects such as pig/zookeeper could then only
pull-in hadoop-core.jar.
Yeah, yeah, this is just to prove that it's weird to have hdfs in 'core' and
not map-reduce! *wink*
bq. Initially one jar: hadoop_mapred.jar
+1
> Refactor src structure, but leave package structure along
> ---------------------------------------------------------
>
> Key: HADOOP-2916
> URL: https://issues.apache.org/jira/browse/HADOOP-2916
> Project: Hadoop Core
> Issue Type: Sub-task
> Reporter: Sanjay Radia
> Fix For: 0.17.0
>
>
> This Jira proposes that the src structure be split as below.
> The package structure remains the same for this Jira. (Package renaming is
> part of other JIras such as HADOOP-2885).
> The idea is that the src will be split BEFORE the package restructuring
> The new proposed src structure is
> src/test - unchanged
> src/java - will no longer exit , its content will be move to one of core,
> hdfs, or mapred
> src/core - this will contain the core classes that hadoop applications need
> to link against.
> It will contain client side libraries of all fs file systems: local,
> hdfs, kfs, etc
> jar name hadoop_core.jar
> src/core/org.apache.hadoop.{conf, fs, filechache, io, ipc, log, metrics,
> net, record, security, tools, util)
> src/core/org.apache.hadoop.dfs - this will contain only the client side
> parts of dfs.
> HADOOP-2885 will rename package dfs to package fs.hdfs
> src/hdfs/org.apache.hadoop.dfs - this will contain only the server side of
> hdfs.
> HADOOP-2885 will rename package dfs to package fs.hdfs later; a
> compatible dfs.DistributedFileSystem will be left for compatibility/
> jar name hadoop_hdfs.jar - this jar can be used to launce NNs and DNs etc.
> src/mapred/org.apache.hadoop.mapred.*
> Initially one jar: hadoop_mapred.jar
> Later this may be split into client-side and server-side jars.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
