hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sudhakara st <sudhakara...@gmail.com>
Subject Re: Question about Yarn
Date Mon, 20 Jan 2014 06:21:01 GMT
Hello Chandler,

Yarn contains separate layers for resource management, scheduling and
map-reduce. Only scheduling and resource management layer is separated to
new daemons , changed and extended in the YARN . The map reduce
functionality and processing data(execution) in parallel framework remain
same as in MR1.

I suggest you to read more on YARN architecture
http://hortonworks.com/hadoop/yarn/
http://blog.cloudera.com/blog/2013/11/migrating-to-mapreduce-2-on-yarn-for-operators/


On Mon, Jan 20, 2014 at 6:42 AM, chandler song <chandler605@gmail.com>wrote:

>
> http://hadoop.apache.org/docs/r2.2.0/hadoop-yarn/hadoop-yarn-site/WritingYarnApplications.html
>
> this one,
>
>
> 2014/1/19 Marco Shaw <marco.shaw@gmail.com>
>
>> Can you clarify?
>>
>> What tutorial and specific sections are you referring to?
>>
>> Marco
>>
>> > On Jan 19, 2014, at 9:26 AM, chandler song <chandler605@gmail.com>
>> wrote:
>> >
>> > hi all
>> >
>> >   I have some question about yarn when I read the tutorial on the
>> website.
>> >
>> >  1) the contain is physical or logic? for example, there are three
>> PCs(A,B,C) on Cluster.  if I allocate one container. it will run on one PC
>> all the time. or the container will work on different PC at different time.
>> >
>> > 2) the contain, I can think it's a virtual PC which can run java
>> application? is my correct?
>> >
>> > 3)about mapreduce, how mapreduce run on yarn? after reading the
>> tutorial, I think yarn and mapreduce is totally different thing. I think
>> the basic unit of yarn is container. map and reduce's basic unit is map and
>> reduce.
>> >
>> > or how yarn handle concurrent? I know in mapreduce, I don't need to
>> think too much about concurrent. because mapreduce will do this for you. it
>> will split data into a small unit and you can do what you do. but I don't
>> find yarn has same thing.
>> >
>>
>
>


-- 

Regards,
...Sudhakara.st

Mime
View raw message