[IMPALA-1062] Impalad Crash When 20 Concurrent Query On The In Memory Cached Hdfs File

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: Impala 1.3.1
• Fix Version/s: Impala 1.4
• Labels: crash impala
• Environment: redhat5.5 , jdk1.7.55
• Assignee: Unassigned
• Reporter: sun tiancheng
• Created: 27/Jun/14 10:42
• Updated: 20/Dec/15 00:05
• Resolved: 28/Jun/14 23:42

Description

When I run 20 concurrent query on impala table,and the file of the table was cached in memory was managered by hdfs cacheadmin, some of the impalad node was crashed.

the crashed node log is :

terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::lock_error> >'
what(): boost::lock_error