Subscribe to Posts by Email

Subscriber Count

    699

Disclaimer

All information is offered in good faith and in the hope that it may be of use for educational purpose and for Database community purpose, but is not guaranteed to be correct, up to date or suitable for any particular purpose. db.geeksinsight.com accepts no liability in respect of this information or its use. This site is independent of and does not represent Oracle Corporation in any way. Oracle does not officially sponsor, approve, or endorse this site or its content and if notify any such I am happy to remove. Product and company names mentioned in this website may be the trademarks of their respective owners and published here for informational purpose only. This is my personal blog. The views expressed on these pages are mine and learnt from other blogs and bloggers and to enhance and support the DBA community and this web blog does not represent the thoughts, intentions, plans or strategies of my current employer nor the Oracle and its affiliates or any other companies. And this website does not offer or take profit for providing these content and this is purely non-profit and for educational purpose only. If you see any issues with Content and copy write issues, I am happy to remove if you notify me. Contact Geek DBA Team, via geeksinsights@gmail.com

Pages

Scripts: Find out a expensive sql statements from AWR Top Events with just three small scripts

A typical day starts like this.

1) Application team complained about slowness in their job

2) DBA looking into this

No clues or details from App team saying this particular sql is that we are running but the response time of the database server is bit slow, so obvious that app team reported slowness.

This time I have chosen event based troubleshooting approach from AWR whether to see am I able to drill down the issue.

What I started first is, lets take a closer look at top classes in the database that is at the moment.

 

set lines 1000
set pages 1000
select wait_class_id, wait_class, count(*) cnt
from dba_hist_active_sess_history
where snap_id between 5205 and 5208
group by wait_class_id, wait_class
order by 3;

  WAIT_CLASS_ID WAIT_CLASS                CNT
------------- ------------------ ----------
    3875070507 Concurrency                 3
3290255840 Configuration               6
3386400367 Commit                     14
2000153315 Network                    15
4217450380 Application                19
4108307767 System I/O                 39
1893977003 Other                     136
1740759767 User I/O                  302
3871361733 Cluster                   786

Note: replace your snap ID’s with your’s

BTW, this is the RAC system and the top Class showing is Cluster (786 count),

Let me see what events are reporting for Wait Class Cluster.

SQL> select event_id, event, count(*) cnt from dba_hist_active_sess_history
where snap_id between 5205 and 5208 and wait_class_id=3871361733
group by event_id, event
order by 3;

  EVENT_ID EVENT                                  CNT
---------- ------------------------------ ----------
2277737081 gc current grant busy                   1
2685450749 gc current grant 2-way                  1
3201690383 gc cr grant 2-way                       1
2701629120 gc current block busy                   3
111015833 gc current block 2-way                  5
3151901526 gc cr block lost                        9
512320954 gc cr request                           9
3905407295 gc current request                     21
1912606394 gc buffer busy acquire                332
661121159 gc cr multi block request             404

 

Well, its gc cr multi block request, nothing but a db file scattered read in other words for a single instance database and the other one gc buffer busy acquire is something that relates to locking or getting latches in the buffer waiting so.

(I am not going through explaining those events since that may fairly require another post)

Lets drill down further what is causing this wait that is sql etc.

SQL> select sql_id, count(*) cnt from dba_hist_active_sess_history
where snap_id between 5205 and 5208
and event_id in ( 1912606394,661121159 )
group by sql_id
having count(*)> 100
order by 2
SQL_ID                    CNT
-------------          ----------
1221u5awzqrb3             288

Note: I have filtered the event_id relates to my events and count > 100, you can change it if you need so. May the count of appearence can be also 1 time but can cause higher wait time for cluster class. So it depends.

Its time to look at the statement for that SQLID I have got above.

 

SQL> select sql_text from v$sql where sql_id='1456sjks32zqrb3';

SQL_TEXT
--------------------------------------------------------------------------------
Select max(sale_dt) from trade

 

Finally I have got the statement which is causing highest waits in my database, and rest of the story as usual , get back to application team ask them whether this statement is related to their job that is reported running slow.  To resolve the same would be another blog story and the above is just meant to identify the statement from top events for a given snap ID’s rather than running a long awr report or ash report when you basically looking for a simple thing like above.

Hope this helps!!!!

Note:- There are many ways to ascertain this information from AWR which can be very easy than above. Like Tanel Poder’s session snapper provides run time information too from ASH.

Comments are closed.