[ https://issues.apache.org/jira/browse/TEZ-3133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15158904#comment-15158904 ]
Hitesh Shah commented on TEZ-3133: ---------------------------------- Feel free to attach the yarn application logs for your job to this jira which will help in the discussion on the mailing list. > many tez tasks only run on one node in cluster > ---------------------------------------------- > > Key: TEZ-3133 > URL: https://issues.apache.org/jira/browse/TEZ-3133 > Project: Apache Tez > Issue Type: Test > Environment: hadoop2.6.0+tez0.7.0 > Reporter: Lvpenglin > > I tested the tez example OrderedWordCount on my cluster.My input file more > than 200MB,and it will generate 4 tasks.But I found this 4 tasks all run on > one node which is different everytime.My cluster contains one master and two > slaves.My idea is this task should be run on different node,for example,there > are two Tokenizer tasks should be run on two different nodes.this can show > the distribution on the cluster.I don't know how to explain this > phenomenon.and I guess if need to modify some configuration or set some > information to realize my imaginative distribution. Who can explain why or > tell me if need some configuration.Thank you very much -- This message was sent by Atlassian JIRA (v6.3.4#6332)