Re: Free SQL Server licenses for Oracle customers

From: Mohamed Houri <mohamed.houri_at_gmail.com>
Date: Sun, 13 Mar 2016 09:05:54 +0100
Message-ID: <CAJu8R6i0Hj=Eu3jVzvj4r0e0BwoAMOrfQ_kf3JC6SwVr9YZ4-Q_at_mail.gmail.com>



At one customer site where I was doing *Oracle* Performance and Tuning consultancy I have been asked to do trouble shoot a *MS SQL Server* performance issue occurring within a stored procedure. Having no prior experience in MS SQL Sever, here's how I embarked in this issue

Me : when this issue happened the last time?

Customer : yesterday between 4 and 5 PM

Me : great, have you a kind of Active Session History (*ASH*)

Customer : no. We have cumulated wait events whithout any sample time

Me : have you *AWR* report?

Customer : no.

Me : have you a kind of *Real time SQL monitoring*?

Customer : no.

Me : Can we trace the Stored Procedure by mean of 10046 trace event

Customer : Yes. We can use the Profiler but there is no guarantee the performance issue will occurs

I have to confess that when trouble shooting Oracle performance issue using ASH and Real time SQL monitoring I rarely need to use 10046 trace event, and very quickly I can at least identify the root cause of the performance issue. It seems to me that this kind of rapid performance diagnostic is not possible in MS SQL Server. This is without speaking about the absence of *Real time SQL monitoring. *In addition, in contrast to Oracle nicely formatted execution plan given by dbms_xplan API, it is almost impossible to read graphical MS SQL Server execution plan.I was obliged to download and use a free tool named *SQL Sentry* in order to have a readable execution plan.

When I was trying to figure out the best way to trouble shoot the above mentioned SQL Server issue, coincidentally at the same time at the same customer Oracle and overnight batch (a Stored Procedure) started to deviate from its normal execution time. This procedure was launched every day between 00h00 and 01h00. I looked at gv$active_session_history where I have found a dominant sql_id showing up between 00H00 and 01H00. This sql_id reveals to be an after insert trigger in which there is an update statement. Just by looking to its historical execution plan I have found that this sql_id has changed its execution plan from a “good” plan to a “bad” plan. I have fixed a SQL profile and the next Oracle batch run went very smoothly. All this took me 2 min to identify the culprit sql_id and about 10 to 15 min to fix a SQL profile and deploy it in TEST and PROD.

Simply put, I believe that when it comes to Performance Trouble shooting issue, with ASH and Real time SQL monitoring, Oracle has a big advantage over MS SQL Server

Best regards

Mohamed Houri

2016-03-12 19:23 GMT+01:00 Mladen Gogala <gogala.mladen_at_gmail.com>:

> On 03/12/2016 11:49 AM, Iggy Fernandez wrote:
>
> I agree. If there is a lot of perceived value, customers are willing to
> pay. If there is little perceived value, customers won't come even if the
> product is free. Making something free is a sign of desperation.
>
> Microsoft has been successful with that before. I remember getting Excel 4
> CD for free during the times when Lotus 1-2-3 was the king of the
> spreadsheets. Now, Lotus is gone and there is only Excel to be found on
> Windows.
>
> --
> Mladen Gogala
> Oracle DBA
> Tel: (347) 321-1217
>
>

-- 

Houri Mohamed

Oracle DBA-Developer-Performance & Tuning

Member of Oraworld-team <http://www.oraworld-team.com/>

Visit My         - Blog <http://www.hourim.wordpress.com/>

Let's Connect -
<http://fr.linkedin.com/pub/mohamed-houri/11/329/857/>*Linkedin
Profile <http://fr.linkedin.com/pub/mohamed-houri/11/329/857/>*

My Twitter <https://twitter.com/MohamedHouri>      - MohamedHouri
<https://twitter.com/MohamedHouri>

--
http://www.freelists.org/webpage/oracle-l
Received on Sun Mar 13 2016 - 09:05:54 CET

Original text of this message