尝试在配置单元上运行查询时,出现以下错误:
FATAL [IPC Server handler 14 on 59490] org.apache.hadoop.mapred.TaskAttemptListenerImpl: Task: attempt_1457495991188_0166_m_000000_3 - exited : Java heap space
查询是:(变量$在执行前被替换)
FROM events_tmp
INSERT INTO TABLE access PARTITION(day='$TODAY')
SELECT $CFIELDS WHERE f0='access' AND f11 LIKE '$TODAY%'
INSERT INTO TABLE access PARTITION(day='$YSTD')
SELECT $CFIELDS WHERE f0='access' AND f11 LIKE '$YSTD%'
INSERT INTO TABLE access PARTITION(day='$TMR')
SELECT $CFIELDS WHERE f0='access' AND f11 LIKE '$TMR%'
INSERT INTO TABLE downloads PARTITION(day='$TODAY')
SELECT $IFIELDS WHERE f0='download' AND f36 LIKE '%dl%' AND f5 LIKE '$TODAY%'
INSERT INTO TABLE downloads PARTITION(day='$YSTD')
SELECT $IFIELDS WHERE f0='download' AND f36 LIKE '%dl%' AND f5 LIKE '$YSTD%'
INSERT INTO TABLE downloads PARTITION(day='$TMR')
SELECT $IFIELDS WHERE f0='download' AND f36 LIKE '%dl%' AND f5 LIKE '$TMR%'
我在/usr/local/hive/conf/hive-env.sh中添加了export hadoop\u heapsize=4096,但它没有改变任何东西。。。。服务器上也有大量的空闲ram。
1条答案
按热度按时间kcwpcxri1#
为了避免这种错误,您需要在mapred-site.xml文件中更新以下与mapreduce堆大小相关的属性。
试试看,会有帮助的。