平时在查看数据库的问题时，会有种迷茫的感觉，如果没有任何人反馈问题，基本上没有主动查找问题的方向，awr,ash都是在得知问题发生的时间段或者一些时间戳来从历史记录中查找相关的信息，个人整理了如下的脚本，能够显示当天的时间段内数据库的负载信息，能够很好掌握数据库的忙闲情况。

来看一个简单的例子，比如我要查看早上6点到中午12点数据库的负载情况

脚本 showdbtime.sh 显示的是在制定的时间段内的 数据库负载的一个综合值。比如6点到7点个小时(60分钟),dbtime如果是120分钟，那么显示的负载就是200%

BEGIN_TIME                                       END_TIME                                  ELAPSED_TIME       DBTIME      WORKLOAD_PER 
----------------------------------- ----------------------------------- ------------- ---------- -------------------- 
201 ** 21-MAY-14 06.07.33.893 PM    201 ** 21-MAY-14 07.07.33.893 PM             60          120            200%

脚本内容如下:

sqlplus -s $DB_CONN_STR@$SH_DB_SID <

set linesize 200 
col begin_time format a35 
col end_time format a35 
col elapsed_time format 99999999.999 
col workload_per format a20 
SELECT 
begin_time,end_time, 
elapsed_time, 
dbtime, 
trunc(dbtime/decode(elapsed_time,0,1,elapsed_time),2)*100||'%' workload_per 
from 
( 
select 
B.SNAP_ID||' ** '||B.END_INTERVAL_TIME begin_time, 
E.SNAP_ID||' ** '||E.END_INTERVAL_TIME end_time, 
EXTRACT(DAY FROM E.END_INTERVAL_TIME - B.END_INTERVAL_TIME) * 1440 + 
EXTRACT(HOUR FROM E.END_INTERVAL_TIME - B.END_INTERVAL_TIME)* 60 + 
EXTRACT(MINUTE FROM E.END_INTERVAL_TIME -B.END_INTERVAL_TIME) + 
EXTRACT(SECOND FROM E.END_INTERVAL_TIME -B.END_INTERVAL_TIME) / 60  elapsed_time, 
db_time.dbtime 
FROM DBA_HIST_SNAPSHOT B, DBA_HIST_SNAPSHOT E, 
                        ( 
                        SELECT b.snap_id begin_snap, e.snap_id end_snap , 
                        round((sum(e.value) - 
                        sum(b.value)) / 1000000 /60,2) dbtime 
                        FROM DBA_HIST_SYS_TIME_MODEL e, DBA_HIST_SYS_TIME_MODEL b, 
                                        ( 
                                        select min(snap_id)  begin_snap,max(snap_id)  end_snap from DBA_HIST_SNAPSHOT where begin_interval_time > sysdate -1 
                                        and EXTRACT(HOUR FROM END_INTERVAL_TIME) between $1 and $2 
                                        ) temp_snap 
                        WHERE 
                         e.STAT_NAME = 'DB time' 
                         and b.snap_id=temp_snap.begin_snap 
                        and e.snap_id =temp_snap.end_snap 
                        AND b.STAT_NAME = 'DB time' 
                        group by e.snap_id,b.snap_id 
                        ) db_time 
WHERE b.begin_interval_time > sysdate -1 
and EXTRACT(HOUR FROM e.END_INTERVAL_TIME) between $1  and $2 
and b.snap_id=db_time.begin_snap 
and e.snap_id=db_time.end_snap 
) 
/

EOF 
exit

在此基础上如果要查看每个小时的数据库负载情况，可以略做一些改进。

脚本showdbtimerpt.sh

sqlplus -s $DB_CONN_STR@$SH_DB_SID < prompt ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 
set head off 
set pages0 
set feedback off 
set serveroutput on 
spool showdbtimerpt_tmp.sh 
begin 
for i in $1..$2 loop 
dbms_output.put_line('ksh showdbtime  '||i||' '||(i+1)); 
end loop; 
end; 
/ 
spool off;

EOF 
clear 
echo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 
echo BEGIN_TIME------------------------- END_TIME--------------------------- ELAPSED_TIME- BTIME----- WORKLOAD_PER-------- 
echo ----------------------------------- ----------------------------------- ------------- ---------- -------------------- 
ksh showdbtimerpt_tmp.sh 
rm showdbtimerpt_tmp.sh 
exit

运行结果类似下面的样子。

BEGIN_TIME                                       END_TIME                                  ELAPSED_TIME       DBTIME      WORKLOAD_PER 
----------------------------------- ----------------------------------- ------------- ---------- -------------------- 
201 ** 21-MAY-14 06.07.33.893 PM    201 ** 21-MAY-14 07.07.33.893 PM             60          120            200%

201 ** 21-MAY-14 07.07.33.893 PM    201 ** 21-MAY-14 08.07.33.893 PM             60          150            250%

201 ** 21-MAY-14 08.07.33.893 PM    201 ** 21-MAY-14 09.07.33.893 PM             60          240            400%

201 ** 21-MAY-14 09.07.33.893 PM    201 ** 21-MAY-14 10.07.33.893 PM             60          60              100%

201 ** 21-MAY-14 10.07.33.893 PM    201 ** 21-MAY-14 11.07.33.893 PM             60          120            200%

可以生成每个时间段的负载报表，这样就一目了然了。可以有针对性的根据负载进行相关的性能抓取。