[SPARK-13115] RandomForest is stuck at computing same stage over and over

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Invalid
• Affects Version/s: 1.5.2
• Component/s: ML, MLlib, Spark Core
• Assignee: Unassigned
• Reporter: Rahul Tanwani
• Created: 01/Feb/16 18:42
• Updated: 02/Feb/16 12:09
• Resolved: 02/Feb/16 03:09

Description

While running the RandomForest regression, the algorithm keeps computing the same stage and does not proceed any further. I have observed the same stage being computed for more than 11 hours. Attached are some of the captures from Spark WebUI.

Also, the spark event logs for this model run could be fetched from Spark Event Logs (https://s3.amazonaws.com/com.tookitaki.public.logs/spark-event-logs). I am running spark-1.5.2 in the standalone local mode. Also, I wanted to know why any stage is marked skipped?

Let me know if you would need more information.