0

I have problem running the given example SimpleShortestPathsComputation with the tiny_graph.txt. I always get stuck at the process of Reduce Part after the Map part. It always shows me 100% maps and 0% reduced and hangs there all the time. Having a hard time looking for solutions in the Web. Does anyone has come across this problem? By the way, my hadoop version is 1.2.1 while giraph is 1.1.0. Thank you!

Alfred
  • 59
  • 1
  • 1
  • 2
  • did you look at the log files? would help finding the issue – peter Nov 10 '14 at 22:08
  • thank you @peter for your reply. as i am very new to this whole hadoop and giraph, how can i look at the log files? it just keep hanging there. i understand that giraph is a map job in hadoop and it keeps hanging there after "INFO mapred.JobClient: map 100% reduce 0%". i tried using hadoop v0.20.203.0 with giraph 1.1.0 but it is still the same outcome. i'm afraid i had configured wrongly. – Alfred Nov 14 '14 at 14:35
  • wait a second. something changes. finally i saw something different. right now its showing me this: 14/11/14 22:27:20 INFO mapred.JobClient: map 100% reduce 0% 14/11/14 22:36:37 INFO mapred.JobClient: map 50% reduce 0% Could someone explain to me what is happening? is the giraph still running? why does it take so long to run? Thank you. – Alfred Nov 14 '14 at 14:39

1 Answers1

0

The problem is that Giraph doesn't have a reducer so basically the reducer never executes. Moreover, the process terminates when only all vertices Do VOTE TO HALT, which never happens.

Here is a quick patch for it. https://issues.apache.org/jira/browse/GIRAPH-756

By using this you can use the (zk node) which is Apache Zookeeper node to halt the process.

Rishabh Wadhawan
  • 31
  • 3