[ https://issues.apache.org/jira/browse/TEZ-1673?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14175292#comment-14175292 ]
Hitesh Shah commented on TEZ-1673: ---------------------------------- Looks good except for the no. of heartbeat events change. Are there any events that may be large in size that would cause a concern here? > Increase the default value of allowed failures per node > ------------------------------------------------------- > > Key: TEZ-1673 > URL: https://issues.apache.org/jira/browse/TEZ-1673 > Project: Apache Tez > Issue Type: Improvement > Reporter: Siddharth Seth > Assignee: Siddharth Seth > Attachments: TEZ-1673.1.txt > > > The current number - 3 is something that was inherited from MapReduce. > Since Tez is affected more by a node being marked as bad - where retries > could be triggered several levels up, I think a higher default value would be > better. I'd propose changing this to 10. -- This message was sent by Atlassian JIRA (v6.3.4#6332)