You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@giraph.apache.org by José Luis Larroque <la...@gmail.com> on 2015/03/24 21:30:30 UTC
Reduce stuck at 0% in PageRankBenchmark - Giraph examples
Hi guys, i'm testing my Single Node Cluster Giraph installation using
PageRankBenchmark example, in the following way:
*$HADOOP_HOME/bin/hadoop jar
/usr/local/hadoop/share/hadoop/mapreduce/giraph-examples-1.1.0-for-hadoop-2.4.0-jar-with-dependencies.jar
org.apache.giraph.benchmark.PageRankBenchmark -v -V 1000 -e 1 -s 5 -w 1*
But after's the mappers complete their job, reducer's don't start. My
question is, this is correct?
Because i understand that if a mapper get executed, a reducer has to start
to take the map's output as an input and finish the work. But i google it,
and always PageRankBenchmark (and ShortestPathExample too) ends with Reduce
at 0%. Here it's an example :
http://stackoverflow.com/questions/25084629/classnotfoundexception-running-giraphrunner-on-a-modified-simpleshortestpathsver
So, i don't no if it's ok that in PageRankBenchmark and ShortestPathExample
says that Reduce is at 0%
I'm using hadoop 2.4, with Phadoop_yarn profile.
Hope someone could help me.
Bye!
Jose
Re: Reduce stuck at 0% in PageRankBenchmark - Giraph examples
Posted by José Luis Larroque <la...@gmail.com>.
Hi again, i get an answer for this in Stack Overflow
<http://stackoverflow.com/questions/29243211/pagerankbenchmark-reduce-at-0-giraph-example/30476769#30476769>,
i will leave it here:
*"Giraph following map-only paradigm. In other words, each worker is
associated with a map task. all of computations is performed within just
map task and communication between map-tasks in order to send / receive
messages is done by using of zookeeper. So, it is unlike of traditional
map-reduce programming paradigm which map outputs transmitted to reducers.
Therefore, there is no reduce task and no map output as well."*
Let's hope that this answer is useful for other's too.
Bye!
2015-03-24 17:46 GMT-03:00 José Luis Larroque <la...@gmail.com>:
> I'm sorry, here are the mapper's logs, please tell me if something else is
> needed.
>
> Here they go in the attachments
>
> Thanks!
>
> 2015-03-24 17:30 GMT-03:00 José Luis Larroque <la...@gmail.com>:
>
>> Hi guys, i'm testing my Single Node Cluster Giraph installation using
>> PageRankBenchmark example, in the following way:
>>
>> *$HADOOP_HOME/bin/hadoop jar
>> /usr/local/hadoop/share/hadoop/mapreduce/giraph-examples-1.1.0-for-hadoop-2.4.0-jar-with-dependencies.jar
>> org.apache.giraph.benchmark.PageRankBenchmark -v -V 1000 -e 1 -s 5 -w 1*
>>
>>
>> But after's the mappers complete their job, reducer's don't start. My
>> question is, this is correct?
>>
>>
>> Because i understand that if a mapper get executed, a reducer has to
>> start to take the map's output as an input and finish the work. But i
>> google it, and always PageRankBenchmark (and ShortestPathExample too) ends
>> with Reduce at 0%. Here it's an example :
>> http://stackoverflow.com/questions/25084629/classnotfoundexception-running-giraphrunner-on-a-modified-simpleshortestpathsver
>>
>>
>> So, i don't no if it's ok that in PageRankBenchmark and
>> ShortestPathExample says that Reduce is at 0%
>>
>>
>> I'm using hadoop 2.4, with Phadoop_yarn profile.
>>
>>
>> Hope someone could help me.
>>
>> Bye!
>>
>>
>> Jose
>>
>>
>
Re: Reduce stuck at 0% in PageRankBenchmark - Giraph examples
Posted by José Luis Larroque <la...@gmail.com>.
I'm sorry, here are the mapper's logs, please tell me if something else is
needed.
Here they go in the attachments
Thanks!
2015-03-24 17:30 GMT-03:00 José Luis Larroque <la...@gmail.com>:
> Hi guys, i'm testing my Single Node Cluster Giraph installation using
> PageRankBenchmark example, in the following way:
>
> *$HADOOP_HOME/bin/hadoop jar
> /usr/local/hadoop/share/hadoop/mapreduce/giraph-examples-1.1.0-for-hadoop-2.4.0-jar-with-dependencies.jar
> org.apache.giraph.benchmark.PageRankBenchmark -v -V 1000 -e 1 -s 5 -w 1*
>
>
> But after's the mappers complete their job, reducer's don't start. My
> question is, this is correct?
>
>
> Because i understand that if a mapper get executed, a reducer has to start
> to take the map's output as an input and finish the work. But i google it,
> and always PageRankBenchmark (and ShortestPathExample too) ends with Reduce
> at 0%. Here it's an example :
> http://stackoverflow.com/questions/25084629/classnotfoundexception-running-giraphrunner-on-a-modified-simpleshortestpathsver
>
>
> So, i don't no if it's ok that in PageRankBenchmark and
> ShortestPathExample says that Reduce is at 0%
>
>
> I'm using hadoop 2.4, with Phadoop_yarn profile.
>
>
> Hope someone could help me.
>
> Bye!
>
>
> Jose
>
>