前言
impala集群出错的一次记录和解决方法以及解决思路。
错误记录
错误信息
Memory limit exceeded Cannot perform hash aggregation. Partitioned input data too many times. This could mean there is too much skew in the data or the memory limit is set too low.
Query信息
就是个这么长的Query语句,Query需要join十多张的表,各种的字段。这只是很多sql中的其中一个。
create TABLE test.cp_ag_info ASSELECT a1.id cid, hr_num, position_num, available_po_num, rs_num, auto_filter_num, read_num, see_num, manual_refuse_num, it_num, auto_refuse_num, forward_num, get_rs_po_num, get_read_rs_po_num, get_see_rs_po_num, get_it_rs_po_num
FROM mysql.cp a1
LEFT JOIN (
SELECT cid, COUNT(DISTINCT uid) hr_num
FROM (
SELECT id uid, testid cid
FROM mysql.dan