2012/9/5 Adarsh Sharma <eddy.adarsh@stripped>
> Actually that query is not my concern :
>
> i have a query that is taking so much time :
> Slow Log Output :
> # Overall: 195 total, 16 unique, 0.00 QPS, 0.31x concurrency _____________
> # Time range: 2012-09-01 14:30:01 to 2012-09-04 14:13:46
> # Attribute total min max avg 95% stddev median
> # ============ ======= ======= ======= ======= ======= ======= =======
> # Exec time 80887s 192us 2520s 415s 1732s 612s 80s
> # Lock time 13ms 0 133us 68us 103us 23us 69us
> # Rows sent 430.89k 0 17.58k 2.21k 12.50k 3.96k 49.17
> # Rows examine 32.30M 0 466.46k 169.63k 440.37k 186.02k 117.95k
> # Query size 65.45k 6 577 343.70 563.87 171.06 246.02
>
> In the logs output :
> # Query_time: 488.031783 Lock_time: 0.000041 Rows_sent: 50
> Rows_examined: 471150
> SET timestamp=1346655789;
> SELECT t0.id, t0.app_name, t0.status, t0.run, t0.user_name,
> t0.group_name, t0.created_time, t0.start_time, t0.last_modified_time,
> t0.end_time, t0.external_id FROM WF_1 t0 WHERE t0.bean_type = 'Workflow'
> ORDER BY t0.created_time DESC LIMIT 0, 50;
>
> The table is near about 30 GB and growing day by day.
>
Just out curiosity, is that table too fragmented? 471k rows are quite a
lot, but 488 of query time is insane. Seems you're reading from disk too
much!
>
> Attaching the table definition & indexes output. I have a index on bean
> type column but cann't understand why it
> examined the all rows of table.
>
Where's the table's schema so we can give it a try?
Manu
