[SPARK-3967] Spark applications fail in yarn-cluster mode when the directories configured in yarn.nodemanager.local-dirs are located on different disks/partitions

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.1.0
• Fix Version/s: 1.2.0
• Component/s: YARN
• Assignee: Christophe Préaud
• Reporter: Christophe Préaud
• Created: 16/Oct/14 09:16
• Updated: 15/May/15 13:35
• Resolved: 15/May/15 13:35

Description

Spark applications fail from time to time in yarn-cluster mode (but not in yarn-client mode) when yarn.nodemanager.local-dirs (Hadoop YARN config) is set to a comma-separated list of directories which are located on different disks/partitions.

Steps to reproduce:
1. Set yarn.nodemanager.local-dirs (in yarn-site.xml) to a list of directories located on different partitions (the more you set, the more likely it will be to reproduce the bug):
(...)
<property>
<name>yarn.nodemanager.local-dirs</name>
<value>file:/d1/yarn/local/nm-local-dir,file:/d2/yarn/local/nm-local-dir,file:/d3/yarn/local/nm-local-dir,file:/d4/yarn/local/nm-local-dir,file:/d5/yarn/local/nm-local-dir,file:/d6/yarn/local/nm-local-dir,file:/d7/yarn/local/nm-local-dir</value>
</property>
(...)
2. Launch (several times) an application in yarn-cluster mode, it will fail (apparently randomly) from time to time

Issue Links

• relates to: SPARK-4896 Don't redundantly copy executor dependencies in Utils.fetchFile
• relates to: SPARK-4764 Ensure that files are fetched atomically
• links to: [Github] Pull Request #2848 (ryan-williams)
• links to: [Github] Pull Request #2855 (preaudc)