Question

我正在尝试为我的Hive查询设置G1GC，但没有骰子：

尝试失败：

set mapreduce.map.java.opts='-Xmx9000m -XX:+UseG1GC -XX:+UseNUMA -XX:MaxGCPauseMillis=600 -XX:+UseStringDeduplication';

set mapreduce.map.java.opts="-Xmx9000m -XX:+UseG1GC -XX:+UseNUMA -XX:MaxGCPauseMillis=600 -XX:+UseStringDeduplication";

set mapreduce.map.java.opts=-Xmx9000m -XX:+UseG1GC -XX:+UseNUMA -XX:MaxGCPauseMillis=600 -XX:+UseStringDeduplication;

这些全部导致：

Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 28
2018-09-11 13:28:27,198 Stage-1 map = 0%,  reduce = 0%
2018-09-11 13:28:40,811 Stage-1 map = 100%,  reduce = 100%
Ended Job = job_1536667701083_0014 with errors
Error during job, obtaining debugging information...
Examining task ID: task_1536667701083_0014_m_000000 (and more) from job job_1536667701083_0014

Task with the most failures(4):
-----
Task ID:
  task_1536667701083_0014_m_000000

URL:
  http://ip-172-31-1-189.eu-central-1.compute.internal:8088/taskdetails.jsp?jobid=job_1536667701083_0014&tipid=task_1536667701083_0014_
m_000000
-----
Diagnostic Messages for this Task:
Exception from container-launch.
Container id: container_1536667701083_0014_01_000006
Exit code: 1
Exception message: Usage: java [-options] class [args...]
           (to execute a class)

如何使用MR后端为Hive查询设置G1GC？

0 个答案: