Chapter 10: Monitoring and Tuning the Database

This chapter introduces you to some of the monitoring and tuning operations as performed through Enterprise Manager.

Approximately 1 hour

Topics

This tutorial covers the following topics:

Place the cursor over this icon to load and view all the screenshots for this tutorial. (Caution: This action loads all screenshots simultaneously, so response time may be slow depending on your Internet connection.)

Note: Alternatively, you can place the cursor over an individual icon in the following steps to load and view only the screenshot associated with that step. You can hide an individual screenshot by clicking it.

Before you perform this tutorial, you should:

1.

Complete Chapter 2: Installing Oracle Software and Building the Database OBE

2.

Complete Chapter 3: Getting Started with Oracle Enterprise Manager OBE

3.

Complete Chapter 4: Configuring the Network Environment OBE

4.

Complete Chapter 5: Managing the Oracle Instance OBE

5.

Complete Chapter 6: Managing Database Storage Structures OBE

6.

Complete Chapter 7: Administering Users and Security OBE

7.

Complete Chapter 8: Managing Schema Objects OBE

8.

Complete Chapter 9: Performing Backup and Recovery OBE

Back to Topic List

Proactively Monitoring your Database

Alerts help you monitor your database proactively. Most alerts are notifications when particular metrics thresholds are crossed. For each alert, you can set critical and warning threshold values. These threshold values are meant to be boundary values that when crossed indicate that the system is in an undesirable state.

In this section, you will perform the following tasks:

A Creating a Tablespace and Table with a Specified Treshold
B. Triggering a Tablespace Space Usage Alert
C. Setting up Notifications

Back to Topic List

Creating a Tablespace and Table with a Specified Threshold

First, you create a new tablespace with a 20 MB data file. This tablespace should be locally managed, and use Automatic Segment Space Management (ASSM). You will then create a new table in this new tablespace. This table will have the Enable Row Movement option set to yes to allow for space reclamation in the table. Perform the following:

1.

Click the Server link on the database home page.

 

2.

Click the Tablespaces link.

 

3.

Click the Create button.

 

4.

Enter TBSALERT as the tablespace name and then click Add to define a datafile for the tablespace.

 

5.

Enter tbsalert01.dbf as the datafile name and 20 MB as the filesize. Click Continue.

Click OK to create the tablespace.

 

6.

Select your new tablespace, TBSALERT, and click Edit.

Click Thresholds to specify the space used warning and critical threshold levels.

 

7.

Select Specify Thresholds in the Space Used (%) section. Enter 60 for the Warning%, and the critical threshold level of 68%. Click Apply.

 

8.

You receive an update confirmation message. Click the Database Instance link to go back to the Server property page.

 

9.

Select the Schema tab. Click the Tables link in the Database Objects section.

 

10.

Click Create.

 

11.

Accept the default of Standard, Heap Organized and click Continue.

 

12.

Enter employees1 in the Name field. Specify SYSTEM as the schema and TBSALERT as the tablespace. Click on the Define Using drop-down list and select SQL. The page is refreshed. Enter select * from hr.employees in the CREATE TABLE AS field. Click Options.

 

13.

Select Yes in the Enable Row Movement drop-down menu. Click OK to complete the table creation.

 

14.

Your table has been created. Click the Database Instance breadcrumb.

 

Back to Topic

Triggering a Tablespace Space Usage Alert

You will now update the table to trigger a space utilization alert. Perform the following:

1.

Open a SQL*Plus session and connect as the SYSTEM user as follows:

sqlplus system/<password>

 

 

2.

Copy and paste the following SQL commands into your SQL*Plus session to simulate user activity on the EMPLOYEES1 table:

begin
  for i in 1..1000 loop
    insert into employees1
    select * from hr.employees;
    commit;
  end loop;
end;
/

 

3.

Return to Enterprise Manager and click the Tablespaces link on the Server page.

 

4.

Notice that the TBSALERT tablespace space used percentage has increased.

 

5.

Return to the SQL*Plus window and copy and paste the following commands into your SQL*Plus session to simulate more user activity on the EMPLOYEES1 table:

delete employees1 where department_id = 50;
begin
  for i in 1..500 loop
    insert into employees1
    select * from hr.employees;
    commit;
  end loop;
end;
/

 

6.

Go to the Enterprise Manager window. Refresh your browser (for Linux Mozilla, select View from the menubar then select Reload). Notice that the TBSALERT tablespace space usage percentage has increased.

 

7.

Switch back to the SQL*Plus window and copy and paste the following commands into your SQL*Plus session to simulate more user activity on the EMPLOYEES1 table:

begin
  for i in 1..500 loop
     insert into employees1
     select * from hr.employees;
     commit;
  end loop;
end;
/

 

8.

Copy and paste the following SQL commands into your SQL*Plus session to simulate user activity on the EMPLOYEES1 table:

delete employees1 where department_id = 30;
commit;
delete employees1 where department_id = 100;
commit;
delete employees1 where department_id = 50;
commit;
delete employees1 where department_id = 80;
commit;

exit

 

9.

Go to the Enterprise Manager window. Refresh your browser (for Linux Mozilla, select View from the menubar then select Reload). Notice that the TBSALERT tablespace space usage percentage has now exceeded the critical threshold level of 60%.

 

10.

While you are waiting for the space usage alert to be displayed on the Enterprise Manager home page, review the table segment statistics. Click the Database breadcrumb, the Schema tab and then click the Tables link.

 

11.

To locate the SYSTEM.EMPLOYEES1 table, enter system in the Schema field and emp in the Object Name field. Click Go.

 

12.

Select the EMPLOYEES1 table. Click Edit.

 

13.

Click Segments.

 

14.

Notice the percentage of wasted space in the EMPLOYEES1 table. You may be able to resolve the tablespace space usage alert by reclaiming unused space in this table.

On this same page, you can project the EMPLOYEES1 table's future space usage by specifying a date range for Space Usage Trend and clicking Refresh. Because there has not been enough activity history on the EMPLOYEES1 table, you will not see very meaningful data in the space usage analysis graph. Click the Database tab.

 

15.

Click the browser refresh/reload button a few times. Scroll down to the Space Summary section of the Home page. Locate a red x and a number next to Problem Tablespaces. Scroll down to the Alerts table.

 

16.

You should see a Tablespaces Full alert. Click the Tablespace TBSALERT is 75 percent full link. (Note: Your system may display slightly different data).

 

17.

The Tablespaces page is displayed. Note that recommendations are given on how to resolve the issue by clicking on the Segment Advisor Recommendations button. However, for the purposes of this exercise, do not make any changes at this time. Note: If you do not see the graph, manually refresh the page.

Remain on this page to continue with the next exercise.

 

Back to Topic

Setting Up Notification

You can optionally provide notification when events that require your intervention arise. By default, alerts in critical state such as Database Down, Generic Alert Log Error Stats, and Tablespace Used are set up for notification. Perform the following:

1.

Click Setup at the top of the Database home page.

 

2.

Click Notification Methods.

 

3.

Enter <your mailserver> in the Outgoing Mail (SMTP) Server field, dbaalert in the Identify Sender As field and notify01@oracle.com in the Sender's E-mail Address field and click Apply.

 

4.

Your update was successful. Click Preferences at the top of the page.

 

5.

Click Add Another Row in the E-mail Addresses section.

 

6.

Enter notify01@oracle.com as the email address and click Apply. Then click Database to return to the Database Home page.

 

Back to Topic

Diagnosing and Resolving Performance Problems

At times database performance problems arise that require your diagnosis and correction. Sometimes problems are brought to your attention by users who complain about slow performance. Other times you might notice performance spikes in the Host CPU chart on the home page.

In all cases, these problems are flagged by the Automatic Database Diagnostics Monitor (ADDM), which does a top-down system analysis every 60 minutes by default and reports its findings on the Oracle Enterprise Manager Home page. ADDM runs automatically every 60 minutes to coincide with the snapshots taken by the Automatic Workload Repository (AWR). Its output consists of a description of each problem it has identified, and a recommended action.

A Creating a Performance Finding
B. Resolving the Performance Finding using ADDM

Back to Topic List

Creating a Performance Finding

To show how ADDM works, you will create a performance finding. In this case, you will create a session waiting on a row lock. To perform certain operations such as updates and deletes, the session must obtain a lock on the row. Perform the following steps to create a performance finding:

1.

Open a terminal window and execute the following commands:

sqlplus hr/<password>
create table emp as select * from employees;
delete emp; 

 

2.

Open another terminal window and execute the following commands to create a row locking conflict:

sqlplus hr/<password>
delete emp;

 

3.

Click Performance in your Enterprise Manager window.

 

4.

Click Application in the Average Active Sessions section.

You see that the sessions waiting is very high. Wait about 10 minutes and scroll down to the bottom of the window.

.

 

5.

You will now create a snapshot to capture the performance finding. Click on Snapshots.

 

6.

Click Create to create a snapshot.

 

7.

Click Yes to create a Manual Snapshot.

 

8.

Once the snapshot is created, click the Database tab.

 

9.

Scroll down the page. A performance finding is now detected and displayed as an alert in the ADDM Performance Analysis section.

 

Back to Topic

Resolving the Performance Finding using ADDM

When a performance finding is encountered, you can use ADDM to resolve it. Perform the following:

1.

Click the finding Row lock waits.

 

2.

Click the SQL ID in the Rationale section.

 

3.

Review the information on the SQL Details page. Click the Database Instance breadcrumb.

 

4.

Click Performance. Scroll down and select Blocking Sessions under Additional Monitoring Links.

Note: If Blocking Sessions does not appear in the Additional Monitoring Links, click Home to return to the Database Home page, and then click the Performance tab again.

 

5.

Make sure the highest level HR is selected and click Kill Session.

 

6.

Click Yes to kill the session.

 

7.

The session has been killed. Click the Database tab .

 

8.

Return to your SQL*Plus sessions and exit from each.

 

Back to Topic

Using the SQL Tuning Advisor

Create a directory named $HOME/wkdir. Download the sqltune.tar file and unzip the file into the $HOME/wkdir directory. Perform the following steps:

1.

Connect as SYSDBA in Enterprise Manager Database Control and navigate to the Performance tab of the Database Control Home page. On the Performance page, make sure that the View Data field is set to Real Time: 15 second Refresh.

 

2.

Open a terminal emulator window connected as the oracle user. Change your current directory to your wkdir directory. Then, enter the following command from the OS prompt:

./setup_dina.sh

 

3.

Execute the start_dinas.sh script as follows:

./start_dinas.sh

 

4.

Return to Enterprise Manager and observe the Performance page for six minutes.

 

5.

Return to your Database Home page. You will now determine the problem. If the time corresponding to the problematic time period corresponds with the latest ADDM run detected by Database Control, you should find the link corresponding to the correct performance analysis directly in the ADDM Performance Analysis section of the Database Control home page. If you do not see the ADDM results, return to the Performance page and click the View ADDM Run icon below the Average Active Sessions graph.

Click the finding with the highest impact on the database time. It should correspond to a SQL Tuning recommendation.

 

6.

On the Performance Finding Details page you see the high-load SQL statement captured by the ADDM analysis. The information provided indicates that there will be a significant benefit if you tune this statement. Click the SQL text.

 

7.

The SQL Details page is displayed. Click Schedule SQL Tuning Advisor.

 

8.

The Schedule Advisor page is displayed. Click Submit.

 

9.

The SQL Tuning Advisor task is scheduled. The page will advance when the task completes.

 

10.

Recommendations are displayed. In this case, the recommendation is to create a SQL Profile in order to get a better execution plan. Click Implement.

 

11.

On the Confirmation window, click Yes.

 

12.

The SQL Profile is created.

 

13.

Return to your terminal window. To see the changes you implemented, you must re-execute the SQL. Stop and start the workload by executing the following commands:

./stop_dinas.sh

./start_dinas.sh

 

14.

Return to Enterprise Manager and access the Performance page to see the benefit of your tuning.

 

15.

Return to your terminal window. Clean up your environment by executing the following commands:

./stop_dinas.sh

./cleanup_dina.sh

 

Back to Topic List

Using the SQL Access Advisor

The SQL Access Advisor provides a number of procedures which can be called to help decide which materialized views and indexes to create and drop. It makes this decision using either a hypothetical workload, which it bases on your schema, or from an actual workload which can be provided by the user, from Oracle Trace or from the contents of the SQL cache.

Workloads may also be filtered according to different criteria, such as only use queries containing these tables or queries which have a priority between this range.

A Preparing the Environment
B. Using the SQL Cache to Get Recommendations
C. Reviewing and Implementing the Recommendations

Back to Topic List

Preparing the Environment

To prepare the environment for using the SQL Access Advisor, perform the steps below. Materialized views and indexes can be present when the advisor is run, but for the purposes of this example they are removed so that you can see what the advisor will recommend. You need to also set up the cache so that the SQL Access Advisor can generate recommendations. Perform the following:

1.

Open a terminal window and execute the following commands to clean up your environment:

sqlplus system/<password>

SELECT * FROM user_objects
WHERE object_type = 'MATERIALIZED VIEW';

If any of the following materialized views exist, delete them as shown:

DROP MATERIALIZED VIEW all_cust_sales_mv;
DROP MATERIALIZED VIEW costs_mv;
DROP MATERIALIZED VIEW costs_pm_mv;
DROP MATERIALIZED VIEW cust_sales_mv;
DROP MATERIALIZED VIEW some_cust_sales_mv;
DROP MATERIALIZED VIEW cust_id_sales_aggr;
DROP MATERIALIZED VIEW sales_cube_mv;
DROP MATERIALIZED VIEW sales_gby_mv;
DROP MATERIALIZED VIEW CUST_TOTAL_SALES_MV;
DROP MATERIALIZED VIEW CUST_SALES_TIME_MV;

 

2.

Now you need to create the cache. Ensure that the SH user is unlocked and then execute the following commands:

alter system flush shared_pool;
grant advisor to sh;

connect sh/<password>;
SELECT c.cust_last_name, sum(s.amount_sold) AS dollars,
sum(s.quantity_sold) as quantity
FROM sales s, customers c, products p
WHERE c.cust_id = s.cust_id
AND s.prod_id = p.prod_id
AND c.cust_state_province IN ('Dublin','Galway')
GROUP BY c.cust_last_name;


SELECT c.cust_id, SUM(amount_sold) AS dollar_sales
FROM sales s, customers c WHERE s.cust_id= c.cust_id GROUP BY c.cust_id;
select sum(unit_cost) from costs group by prod_id;

 

 

Back to Topic

Using the SQL Cache to Get Recommendations

You will use the SQL Cache you just set up to obtain recommendations from the SQL Access Advisor. Perform the following:

1.

Scroll to the bottom of the Database Home page and click Advisor Central under Related Links.

 

2.

Click the SQL Advisors link.

 

3.

Click SQL Access Advisor.

 

4.

Select Recommend new access structures and click Continue.

 

5.

Ensure Current and Recent SQL Activity is selected. Expand Filter Options.

Select Filter Workload Based on these Options. Select Include only SQL statements executed by these users. Enter SH in the Users field and click Next.

 

6.

Select Indexes and Materialized Views and click Next.

 

7.

Enter the task name SQLACCESS<Today's Date>, select Standard for the Schedule Type and click Next.

 

8.

At the summary window, click Submit.

 

9.

A confirmation message appears.

 

Back to Topic

Reviewing and Implementing the Recommendations

Now you can look at the results and implement them if you wish. Perform the following:

1.

On the Advisor Central page, ensure your job is selected and click View Result.

 

2.

The Summary page is displayed. Click Recommendations.

Click on the Recommendation ID 1 to see the details of the Recommendations.

 

3.

Here you can customize the Object Name, Schema and Tablespace to implement the recommendations. Scroll down and change the Schema Name for the Create Materialized View to SH and click OK.

 

4.

To implement the recommendations, click Schedule Implementation.

 

7.

Enter SQLACCESSIMPL<today's date> for the Job Name and click Submit.

 

8.

Your implementation job was created and is now executing.

 

9.

Click Database to return to the Database Home page. Click Schema.

 

10.

Click Materialized Views in the Materialized Views section.

 

11.

Enter SH in the schema field and click Go.

 

12.

Notice that the newly created Materialized View appears in the list. Click the Database breadcrumb then click the Home tab.

 

Back to Topic

Enabling Automatic Shared Memory Management Using the Memory Advisor

In this section, you will proactively manage and automate some of the tasks related to Oracle Instance memory configuration. By automating memory configuration, you have more time to deal with real application or business issues that affect your enterprise.

The Memory Advisor is an intelligent expert system within the Oracle database that proactively determines optimal settings for various SGA and PGA components. When automated, Oracle will automatically adjust the settings for the various pools and caches according to the requirements of the workload.

In this section you will verify that automatic shared memory management is enabled. If it is disabled, you can enable it by following the steps outlined below.

1.

Scroll down to the bottom of the home page and click on Advisor Central under Related Links.

 

2.

Select Memory Advisors.

 

3.

Confirm that Automatic Shared Memory Management is enabled. If you would like to disable Automatic Shared Memory Management, click Disable. Click the Database tab to return to the Database Home page.

 

Back to Topic List

In this tutorial, you learned how to:

Create a tablespace with a specified threshold.
Triggering a tablespace usage alert.
Set metric thresholds.
Create a performance finding.
Solve the performance finding using ADDM.

Back to Topic List

Place the cursor over this icon to hide all screenshots.