runstats安裝方法
runstats是TOM老人家寫的一款很好的基準測試小工具,下面簡單介紹其安裝及使用方法。
一、以sys登入進行授權
grant select on sys.v_$timer to suk;
grant select on v_$mystat to suk;
grant select on sys.v_$statname to suk;
grant select on sys.v_$latch to suk;
二 安裝Runstats包
以suk登入,分別執行如下步驟:
1 建立stats檢視
create or replace view stats
as
select 'STAT..' || a.name name, b.value
from v$statname a, v$mystat b
where a.statistic# = b.statistic#
union all
select 'LATCH.' || name, gets
from v$latch;
2 建立run_stats臨時表
create global temporary table run_stats
( runid varchar2(15),
name varchar2(80),
value int
)on commit preserve rows;
3 建立Runstats包頭
create or replace package runstats
as
procedure rs_start;
procedure rs_middle;
procedure rs_stop( p_difference_threshold in number default 0);
end;
p_difference_threshold引數的含義是:前後兩個相同的項的統計結果相差的數目的絕對值大於等於該引數設定的值,該統計項才會被顯示。預設值是0表示所有的結果都顯示。
4 建立Runstats包體
create or replace package body runstats
as
g_start number;
g_run1 number;
g_run2 number;
procedure rs_start
is
begin
delete from run_stats;
insert into run_stats
select 'before', stats.* from stats;
g_start := dbms_utility.get_time;
end;
procedure rs_middle
is
begin
g_run1 := (dbms_utility.get_time-g_start);
insert into run_stats
select 'after 1', stats.* from stats;
g_start := dbms_utility.get_time;
end;
procedure rs_stop(p_difference_threshold in number default 0)
is
begin
g_run2 := (dbms_utility.get_time-g_start);
--add a line here to avoid ora-20000
dbms_output.enable(480000);
dbms_output.put_line
( 'Run1 ran in ' || g_run1 || ' hsecs' );
dbms_output.put_line
( 'Run2 ran in ' || g_run2 || ' hsecs' );
dbms_output.put_line
( 'run 1 ran in ' || round(g_run1/g_run2*100,2) ||
'% of the time' );
dbms_output.put_line( chr(9) );
insert into run_stats
select 'after 2', stats.* from stats;
dbms_output.put_line
( rpad( 'Name', 30 ) || lpad( 'Run1', 12 ) ||
lpad( 'Run2', 12 ) || lpad( 'Diff', 12 ) );
for x in
( select rpad( a.name, 30 ) ||
to_char( b.value-a.value, '999,999,999' ) ||
to_char( c.value-b.value, '999,999,999' ) ||
to_char( ( (c.value-b.value)-(b.value-a.value)), '999,999,999' ) data
from run_stats a, run_stats b, run_stats c
where a.name = b.name
and b.name = c.name
and a.runid = 'before'
and b.runid = 'after 1'
and c.runid = 'after 2'
-- and (c.value-a.value) > 0
and abs( (c.value-b.value) - (b.value-a.value) )
> p_difference_threshold
order by abs( (c.value-b.value)-(b.value-a.value))
) loop
dbms_output.put_line( x.data );
end loop;
dbms_output.put_line( chr(9) );
dbms_output.put_line
( 'Run1 latches total versus runs -- difference and pct' );
dbms_output.put_line
( lpad( 'Run1', 12 ) || lpad( 'Run2', 12 ) ||
lpad( 'Diff', 12 ) || lpad( 'Pct', 10 ) );
for x in
( select to_char( run1, '999,999,999' ) ||
to_char( run2, '999,999,999' ) ||
to_char( diff, '999,999,999' ) ||
to_char( round( run1/run2*100,2 ), '99,999.99' ) || '%' data
from ( select sum(b.value-a.value) run1, sum(c.value-b.value) run2,
sum( (c.value-b.value)-(b.value-a.value)) diff
from run_stats a, run_stats b, run_stats c
where a.name = b.name
and b.name = c.name
and a.runid = 'before'
and b.runid = 'after 1'
and c.runid = 'after 2'
and a.name like 'LATCH%'
)
) loop
dbms_output.put_line( x.data );
end loop;
end;
end;
至此,Runstats包安裝完畢。
三、使用
舉例如下:
--建立測試表
create table x(id number);
SQL> set serverout on
--記錄開始狀態
SQL> exec runstats.rs_start;
PL/SQL procedure successfully completed
SQL>
SQL> Begin
2 for i in 1..1000 Loop
3 insert into x values(i);
4 end loop;
5 end;
6 /
PL/SQL procedure successfully completed
--記錄執行完第一段sql後的狀態
SQL> exec runstats.rs_middle;
PL/SQL procedure successfully completed
SQL>
SQL> Begin
2 for i in 1..1000 Loop
3 Execute Immediate 'insert into x values('||i||')';
4 end loop;
5 end;
6 /
PL/SQL procedure successfully completed
--記錄執行完第二段sql後的狀態,並顯示所有的統計結果
SQL> exec runstats.rs_stop;
Run1 ran in 4333 hsecs
Run2 ran in 5825 hsecs
run 1 ran in 74.39% of the time
Name Run1 Run2 Diff
LATCH.job_queue_processes para 0 1 1
LATCH.kwqit: protect wakeup ti 1 2 1
LATCH.ncodef allocation latch 0 1 1
LATCH.session switching 0 1 1
LATCH.transaction branch alloc 0 1 1
STAT..buffer is not pinned cou 3 2 -1
STAT..cleanout - number of ktu 2 3 1
STAT..cluster key scans 1 0 -1
STAT..index fetch by key 1 0 -1
STAT..sorts (rows) 2,409 2,408 -1
STAT..redo ordering marks 3 2 -1
STAT..no work - consistent rea 3 2 -1
STAT..messages sent 1 0 -1
STAT..free buffer requested 19 20 1
STAT..cluster key scan block g 1 0 -1
STAT..calls to kcmgcs 4 5 1
STAT..active txn count during 2 3 1
LATCH.sort extent pool 0 1 1
LATCH.simulator lru latch 1 2 1
LATCH.cache buffer handles 2 4 2
LATCH.object stats modificatio 2 0 -2
STAT..workarea executions - op 10 8 -2
STAT..sorts (memory) 6 4 -2
STAT..change write time 0 2 2
STAT..calls to kcmgas 4 2 -2
LATCH.active checkpoint queue 16 19 3
LATCH.cache buffers lru chain 25 30 5
LATCH.mostly latch-free SCN 14 19 5
LATCH.lgwr LWN SCN 14 19 5
LATCH.channel operations paren 30 36 6
LATCH.library cache load lock 6 0 -6
LATCH.session timer 14 20 6
LATCH.enqueues 69 76 7
STAT..shared hash latch upgrad 7 0 -7
STAT..index scans kdiixs1 7 0 -7
STAT..deferred (CURRENT) block 7 0 -7
STAT..consistent gets - examin 10 3 -7
STAT..commit cleanouts 9 0 -9
STAT..commit cleanouts success 9 0 -9
STAT..enqueue releases 13 4 -9
LATCH.redo writing 56 67 11
STAT..enqueue requests 15 4 -11
STAT..redo entries 1,508 1,495 -13
LATCH.dml lock allocation 35 49 14
STAT..execute count 1,025 1,011 -14
LATCH.redo allocation 1,554 1,569 15
LATCH.session idle bit 53 38 -15
STAT..calls to get snapshot sc 1,024 1,007 -17
LATCH.session allocation 30 52 22
STAT..bytes received via SQL*N 2,541 2,565 24
LATCH.undo global data 34 61 27
STAT..db block changes 3,047 3,020 -27
STAT..db block gets 1,568 1,541 -27
LATCH.messages 124 160 36
LATCH.enqueue hash chains 121 158 37
STAT..parse time cpu 2 39 37
STAT..parse time elapsed 1 45 44
STAT..CPU used by this session 8 59 51
STAT..CPU used when call start 8 59 51
STAT..recursive cpu usage 2 55 53
LATCH.simulator hash latch 1 136 135
STAT..recursive calls 1,139 1,002 -137
LATCH.checkpoint queue latch 760 904 144
LATCH.SQL memory manager worka 938 1,206 268
STAT..session logical reads 1,598 2,551 953
STAT..consistent gets 30 1,010 980
STAT..parse count (total) 24 1,011 987
STAT..opened cursors cumulativ 18 1,011 993
STAT..parse count (hard) 2 1,001 999
LATCH.cache buffers chains 8,096 10,209 2,113
LATCH.row cache objects 32 4,024 3,992
LATCH.row cache enqueue latch 26 4,024 3,998
STAT..redo size 290,276 285,756 -4,520
LATCH.library cache pin alloca 184 6,162 5,978
LATCH.library cache pin 2,342 9,291 6,949
LATCH.child cursor hash table 18 7,007 6,989
LATCH.library cache 2,589 23,497 20,908
LATCH.shared pool 1,386 26,344 24,958
STAT..session uga memory 65,464 0 -65,464
STAT..session pga memory 65,536 0 -65,536
STAT..session pga memory max 327,680 0 -327,680
Run1 latches total versus runs -- difference and pct
Run1 Run2 Diff Pct
18,588 95,205 76,617 19.52%
--從這裡可以看出,不繫結變數佔用的栓鎖是繫結變數的5倍
PL/SQL procedure successfully completed
--如果只想顯示同一項的統計值前後差異超過1000的項,則執行
SQL> exec runstats.rs_stop(1000);
來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/231499/viewspace-63737/,如需轉載,請註明出處,否則將追究法律責任。
相關文章
- runstats_pkg包的安裝
- Runstats 指令碼指令碼
- Tom的runstats工具
- kazoo安裝方法
- rlwrap 安裝方法
- 安裝sysbench 方法
- db2 reorg,runstatsDB2
- Ubuntu 安裝 SCRAPY 方法Ubuntu
- nvm的安裝方法
- cocoapods安裝方法
- CDH 5.13.0安裝方法
- snmp的安裝方法
- runstats工具的配置及使用
- 轉: Ubuntu 安裝字型方法Ubuntu
- MYSQL完全解除安裝方法MySql
- Kubernetes - 安裝方法
- 選擇一種安裝方法安裝Linux(轉)Linux
- DB2 PL/SQL Example: RunstatsDB2SQL
- 效能比較工具runStats_pkg
- SQL最佳化好助手:RunstatsSQL
- 【實驗】【Tom】runstats工具使用演示
- 如何安裝pdf印表機 安裝pdf印表機的方法
- linux上安裝Docker(非常簡單的安裝方法)LinuxDocker
- windows安裝mysql以及安裝Navicat Premium並破解的方法WindowsMySqlREM
- mysql在linux上cmake安裝方法(自己安裝版)MySqlLinux
- Mac教程——字型的安裝方法Mac
- vmware workstation 14 pro安裝方法
- kubernetes的安裝方法
- python 安裝模組的方法Python
- Centos下安裝mariaDB方法CentOS
- WSL2安裝systemd方法
- IDE介面硬碟安裝方法IDE硬碟
- Linux解除安裝jdk方法LinuxJDK
- Nagiosql安裝的配置方法iOSSQL
- windows10硬碟安裝方法Windows硬碟
- MySQL簡介和安裝方法MySql
- 軟體安裝常用方法(轉)
- Macromedia Flex 安裝註冊方法MacFlex