Seeing higher than expected CPU/MEM usage for the java process started by the
sqlline.py client.
From top:
top - 14:37:32 up 7 days, 22:15, 2 users, load average: 25.52, 9.74, 7.89
Tasks: 509 total, 1 running, 508 sleeping, 0 stopped, 0 zombie
Cpu(s): 61.2%us, 6.3%sy, 0.0%ni, 31.0%id,
Can anyone explain why the client would be burning so much CPU and memory if
the result is a single row?
I suspect we configured something wrong on Phoenix but we are having a hard
time figuring it out.
Thanks in advance.
Regards,
John
> On Aug 24, 2016, at 9:54 AM, Aaron Molitor wrote:
The client typically does a merge sort when combining results from
paralllel scans. Not sure if this would explain the CPU/memory usage you're
seeing. Can you narrow it down to a particular query that's causing the
issue?
Thanks,
James
On Thu, Aug 25, 2016 at 6:49 AM, John Leach
wrote:
> Can an
Yeah, this query.
QUERY:
SELECT SUM(L_EXTENDEDPRICE) / 7.0 AS AVG_YEARLY
FROM
TPCH.LINEITEM,
TPCH.PART
WHERE
P_PARTKEY = L_PARTKEY
AND P_BRAND = 'BRAND#23'
AND P_CONTAINER = 'MED BOX'
AND L_QUANTITY < (
SELECT 0.2 * AVG(L_QUANTITY)
FROM
TPCH.LINEITEM
WHERE
L_PARTKEY = P_PA
Hi John,
Would you mind sharing the query plan for this query (by running "EXPLAIN
")?
Thanks,
Maryann
On Thu, Aug 25, 2016 at 11:19 AM, John Leach wrote:
> Yeah, this query.
>
> QUERY:
> SELECT SUM(L_EXTENDEDPRICE) / 7.0 AS AVG_YEARLY
> FROM
> TPCH.LINEITEM,
> TPCH.PART
> WHERE
> P_PARTKE
1/1 EXPLAIN SELECT SUM(L_EXTENDEDPRICE) / 7.0 AS AVG_YEARLY
FROM
TPCH.LINEITEM,
TPCH.PART
WHERE
P_PARTKEY = L_PARTKEY
AND P_BRAND = 'BRAND#23'
AND P_CONTAINER = 'MED BOX'
AND L_QUANTITY < (
SELECT 0.2 * AVG(L_QUANTITY)
FROM
TPCH.LINEITEM
WHERE
L_PARTKEY = P_PARTKEY
)
;
+