Issue Details (XML | Word | Printable)

Key: MAPREDUCE-231
Type: Task Task
Status: Open Open
Priority: Major Major
Assignee: Owen O'Malley
Reporter: Owen O'Malley
Votes: 0
Watchers: 17
Operations

If you were logged in you would be able to see more operations.
Hadoop Map/Reduce

Split map/reduce into sub-project

Created: 19/Nov/08 06:48 PM   Updated: 22/Sep/09 06:12 AM
Component/s: None
Affects Version/s: None
Fix Version/s: None

Time Tracking:
Issue & Sub-Tasks
Issue Only
Not Specified

Issue Links:
Blocker
 

Sub-Tasks  All   Open   

 All   Comments   Work Log   Change History   Subversion Commits      Sort Order: Ascending order - Click to sort in descending order
Nigel Daley added a comment - 19/May/09 04:28 AM
Owen, thoughts on how we should rename these jar files in the mapred and hdfs projects?

hadoop-0.21.0-dev-ant.jar
hadoop-0.21.0-dev-core.jar
hadoop-0.21.0-dev-examples.jar
hadoop-0.21.0-dev-test.jar
hadoop-0.21.0-dev-tools.jar


Giridharan Kesavan added a comment - 20/May/09 06:44 AM
With this split I'm also thinking about changing the folder structure to follow maven standards. Any comments?
tnx!

Nigel Daley added a comment - 20/May/09 06:51 AM
What would those standards be?

dhruba borthakur added a comment - 20/May/09 07:04 AM
I am assuming that HadoopQA will run all unit tests (both dfs and map-reduce) when a patch is submitted for the dfs sub-project (or for the map-reduce subproject).

Giridharan Kesavan added a comment - 20/May/09 07:39 AM

What would those standards be?

java sources would go in here.
src/main/java
test sources would go in here
src/test/java
For details : http://maven.apache.org/guides/introduction/introduction-to-the-standard-directory-layout.html


Tom White added a comment - 20/May/09 10:25 AM
As far as jar naming goes, how about

hadoop-{core,hdfs,mapred}-0.21.0-dev.jar
hadoop-{core,hdfs,mapred}-{ant,examples,test,tools}-0.21.0-dev.jar

Having the version number at the end seems to be the standard way to name things in the Ivy/Maven world.


Nigel Daley added a comment - 20/May/09 02:49 PM
+1 on maven layout.

+1 on Tom's jar name suggestions.

Regarding jar names, we'll need to update bin scripts. Here are current references in the top-level bin scripts:

hadoop-config.sh:for f in $HADOOP_CORE_HOME/hadoop-*-core.jar; do
hadoop-config.sh:for f in $HADOOP_CORE_HOME/hadoop-*-tools.jar; do
hadoop-config.sh:for f in $HADOOP_CORE_HOME/build/hadoop-*-tools.jar; do
hadoop-config.sh:  for f in $HADOOP_HDFS_HOME/hadoop-*-hdfs.jar; do
hdfs:for f in $HADOOP_HDFS_HOME/hadoop-*-hdfs.jar; do
mapred:for f in $HADOOP_MAPRED_HOME/hadoop-*-mapred.jar; do
rcc:for f in $HADOOP_HOME/hadoop-*-core.jar; do

Contrib bin scripts, Makefiles, etc also need to be checked and updated. Owen, this seems beyond what Giri should have to do for getting the build working. Should we file a separate Jira for that once the names are settled on.


Iyappan Srinivasan added a comment - 22/Sep/09 06:12 AM
Please look at MAPREDUCE-1004, ant binary not copying jar files properly.