chukwa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jerome Boulon (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CHUKWA-481) Improve demux reducer partitioning algorithm
Date Tue, 27 Apr 2010 23:33:32 GMT

    [ https://issues.apache.org/jira/browse/CHUKWA-481?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12861617#action_12861617
] 

Jerome Boulon commented on CHUKWA-481:
--------------------------------------

Partitioning is the key of M/R so reducing the partitioning function to 2 implementations
will not make sense for everybody. 
I understand that you are interested in case#1 and case#1 will be only good when you can predict
what kind of data you're going to have and to do your grouping function but this will not
be useful for Hive output for example. The ideal case will be to support partitioning function
at the dataType level so everyone can define the partitioning function that is the right for
a specific dataType... but that the ideal case, the minimum will be to have the partitioning
class define in chukwa-demux-conf.xml. This way anybody will be free to implement/configure
the system to match their requirements.



> Improve demux reducer partitioning algorithm
> --------------------------------------------
>
>                 Key: CHUKWA-481
>                 URL: https://issues.apache.org/jira/browse/CHUKWA-481
>             Project: Hadoop Chukwa
>          Issue Type: Improvement
>          Components: MR Data Processors
>         Environment: Redhat EL 5.1, Java 6
>            Reporter: Eric Yang
>            Assignee: Eric Yang
>
> Reducer partitioning for demux could be redefined to optimize for two different use case:
> Case #1, demux is responsible for crunching large volumes of the same data type (dozen
of types).  It will probably make more sense to partition the reducer by time grouping + data
type (extend TotalOrderPartitioner).  I.e. A user can have evenly distributed workload for
each reducer base on time interval.  A distributed hash table like Hbase/voldermort could
be the down stream system to store/cache the data for data serving.  This model is great for
collecting fixed time interval logs like hadoop metrics, and ExecAdaptor which generates repetitive
time series summary.
>  
> Case #2, demux is responsible for crunching hundred of different data type, but small
volumn for each data type.  The current demux implementation is using this model, where a
single data type is reduced by one reducer slot (ChukwaRecordPartitioner).  One draw back
from this model,the data from each data type must have similar volume.  Otherwise, the largest
data volume type becomes the long tail of the mapreduce job.  Materialized report is easy
to generate by using this model because the single reducer per data type has view to all data
of the given demux run.  This model works great for many different application and all logging
through Chukwa Log4j appender.  I.e. web crawl, or log file indexing / viewing.
>  
> I am thinking to change the default Chukwa demux implementation to case #1, and restructure
the current demux as Archive Organizer.  Any suggestion or objection?

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message