我试图在AWS EMR中使用集群（1个主节点，4个核心节点[每个11.25GB]）在hive中运行作业，即时加入（MAP加入）两个表，其中一个表有30万个条目（~11mb），另一个表有近700万条目（注意大表应该是最后一个）。因为它是映射连接，所以第一个小表在分布式缓存中进行散列，在散列了一些条目之后我得到了这个错误。
2015-10-07 11:21:35 Processing rows: 10600000 Hashtable size: 10599999 Memory usage: 5465925664 percentage: 0.143
2015-10-07 11:22:09 Processing rows: 10700000 Hashtable size: 10699999 Memory usage: 6078125280 percentage: 0.159
2015-10-07 11:22:42 Processing rows: 10800000 Hashtable size: 10799999 Memory usage: 2698310976 percentage: 0.071
2015-10-07 11:23:16 Processing rows: 10900000 Hashtable size: 10899999 Memory usage: 3272893216 percentage: 0.086
2015-10-07 11:23:49 Processing rows: 11000000 Hashtable size: 10999999 Memory usage: 3821330720 percentage: 0.10
2015-10-07 11:24:23 Processing rows: 11100000 Hashtable size: 11099999 Memory usage: 4398775280 percentage: 0.115
2015-10-07 11:24:57 Processing rows: 11200000 Hashtable size: 11199999 Memory usage: 5020884784 percentage: 0.132
2015-10-07 11:25:30 Processing rows: 11300000 Hashtable size: 11299999 Memory usage: 5608609488 percentage: 0.147
2015-10-07 11:26:04 Processing rows: 11400000 Hashtable size: 11399999 Memory usage: 6146767072 percentage: 0.161
2015-10-07 11:26:38 Processing rows: 11500000 Hashtable size: 11499999 Memory usage: 2622476744 percentage: 0.069
2015-10-07 11:27:11 Processing rows: 11600000 Hashtable size: 11599999 Memory usage: 3224517304 percentage: 0.084
2015-10-07 11:27:44 Processing rows: 11700000 Hashtable size: 11699999 Memory usage: 3793226384 percentage: 0.099
Execution failed with exit status: 137
Obtaining error information
Task failed!
Task ID:
Stage-4
Logs:
/mnt/var/log/apps/hive.log
FAILED: Execution Error, return code 137 from org.apache.hadoop.hive.ql.exec.mr.MapredLocalTask