I have created Impala PARQUET TABLE OF 50GB data. I am running a query with joins of 5 tables, grouping and sorting of data.
at the end of the query the server did not respond any output data nether any exception is thrown by Impala.
On cloudera manager i can see one node is in bad health few min back.
Query:
select
d_year, s_city, p_brand1,
sum(lo_revenue - lo_supplycost) as profit
from
lineorder_kylin_parqt_ext, dates_kylin_parqt_ext ,customer_kylin_parqt_ext, supplier_kylin_parqt_ext, part_kylin_parqt_ext
where
lo_custkey = c_custkey
and lo_suppkey = s_suppkey
and lo_partkey = p_partkey
and lo_orderdate = d_datekey
and c_region = 'AMERICA'
and s_nation = 'UNITED STATES'
and (d_year = 1997 or d_year = 1998)
and p_category = 'MFGR#14'
group by
d_year, s_city, p_brand1
order by
d_year, s_city, p_brand1;
I got the answer OS was killing impala demon on the node as it was eating memory. I confirmed that from Kernel logs and on Cloudera Manager as well the demon where query was executing found dead.