Skip navigation.

Feed aggregator

Security via policies

Yann Neuhaus - Tue, 2015-02-10 10:15


Few weeks ago, I presented the session on security via Policies for "Les journées SQL Server 2014", organized by the French SQL Server User Group (GUSS) in Paris.


Data Integration Tips: ODI – Use Scenarios in packages

Rittman Mead Consulting - Tue, 2015-02-10 07:16


Here is another question I often hear during ODI Bootcamp or I often read in the ODI Space on OTN :
In a package, should I directly use mappings (or interfaces prior to 12c) or is it better to use the scenarios generated from these mappings?

An ODI package is a workflow used to sequence the different steps of execution of a data integration process. Some of these steps might change the value of a variable, evaluate a variable, perform an administrative task like sending an email or retrieving a file from an FTP. But the core thing is the execution of a mapping and this can be done in two different ways.


Direct Use

It is possible to directly drag and drop a mapping in the package.

Mapping in a Package

In this example the mapping M_FACT_SALES_10 is directly executed within the package, in the same session. This means that the value of the variables will be the same as the ones in the package.

Execution of a package holding a mapping


If there are several mappings, they will be executed sequentially, there is no parallelism here. (Damn, I will never get the double L at the right place in that word… Thanks autocorrect!)

The code for the mapping is generated using the current state of the mapping in the work repository, including any changes done since the package was created. But if a scenario is created for the package PCK_FACT_LOAD_MAP, it will need to be regenerated – or a new version should be generated – to take any change of M_FACT_SALES_10 into account.

One good point for this approach is that we can build a data integration process that will be entirely rolled back in case of failure. If one of the step fails, the data will be exactly as it was before the execution started. This can be done by using the same transaction for all the mappings and disable the commit option for all of them but the last one. So it’s only when everything succeeds that the changes will be committed and available for everyone. Actually it’s not even required to commit on the last step as ODI will anyway issue a commit at the end of a successful session. This technique works of course only if you don’t have any DDL statement on the target tables. You will find more information about it on this nice blog post from Bhabani Rajan : Transaction Control in ODI. Thanks to David Allan for reminding me this!


Using Scenarios

But you can also generate a scenario and then drag and drop it in a packageScenario in a package

Sessions and variables

In this case, the execution of the mapping will not be done within the package session. Instead, on step of this session will use the OdiStartScen command to trigger a new session to execute the scenario.

Execution of a package holding a scenario

We can see here that the second step of the package session (401) got only one task which will run the command to start a new session. The only step in the new session (402) is the mapping and it has the three same tasks as the previous example. Thanks to the fact this is a different session, you could choose to execute it using another context or agent. My colleague Michael brought an interesting use case to me. When one step of your package must extract data from a file on a file server that has its own standalone agent installed to avoid permission issues, you could specify that agent in the Scenario step of your package. So the session for that scenario will be started on that standalone agent while all the other sessions will use the execution agent of the parent session.


But what about variables configured to keep no history? As this is a different session, the values are lost outside of the scope of one session. We therefore need to pass it as startup parameters to the scenario of the mapping.

Passing variables as startup parameters To do this, I go to the Command tab on the scenario step and I add my variables there with the syntax


where <VALUE_TO_ASSIGN> can be the value of variable in the current session itself (401 in our case).

 Code executed

The scenario step name was originally “Execution of the Scenario M_FACT_SALES_10 version 001″ and I slightly modified it to remove the mention of a version number. As I want to always execute the last version of the scenario, I also changed the version in the parameters and set it to -1 instead of 001. There is a small bug in the user interface so if you click somewhere else, it will change back to 001. The workaround is to press the Return key (or Enter) after typing your value.

So by using a scenario, you can choose which version of a scenario you want to execute. Either a specific version (e.g. 003) or the last one, using the value -1. It also means that if you won’t break anything if the package is executed while you are working on the mapping. The package will still use the frozen code of the scenario even if you changed the mapping.

If you create a scenario for the package PCK_FACT_LOAD_SCEN, there is no need to regenerate it when a new scenario is (re-)generated for the mapping. By using -1, it will reference the newly generated scenario.

 Asynchronous Execution

Another advantage of using scenarios is that it supports parallelism (Yes, I got it right!!).

Asynchronous scenario executions

Here I set the “Synchronous / Asynchronous” option to “Asynchronous Mode” for my two scenarios so the two sessions will start at the same time. By adding an OdiWaitForChildSessions step, I can wait for the end of all the sessions before doing anything else. This step will also define in which case you want to report an error. By default, it’s when all the child sessions are in error. But I usually change to parameter to zero, so any failure will cause my package execution to fail as well.


Just a short note, be careful when using the method getPrevStepLog() from the substitution API in a step after executing a scenario. That method will retrieve information about the previous step in the current Session. So about the ODI command execution OdiStartScen, and not about the execution of the scenario itself.



Here is a small recap table to compare both approach :

comparison table

In conclusion, development is generally more robust when using scenarios in packages. There is more control on what is executed and on the parallelism. The only good reason to use mappings directly is to keep everything within the same transaction.

Keep also in mind that Load Plans might be a better alternative than packages with better parallelism, error handling and restartability. Unless you need loops or a persistent session… A comparison between Load Plans and Packages can be found in the Oracle blog post introducing Load Plans.


More tips and core principles demystification coming soon so stay tuned. And follow @mRainey, @markrittman and myself – @JeromeFr – on twitter to be informed of anything happening in the ODI world.

Categories: BI & Warehousing

Some changes to be aware of, as Oracle Application Express 5 nears...

Joel Kallman - Tue, 2015-02-10 06:25
As the release of Oracle Application Express 5 gets closer, I thought it's worth pointing out some changes that customers should be aware of, and how an upgrade to Oracle Application Express 5 could impact their existing applications.

  1. As Trent Schafer (@trentschafer) noted in his latest blog post, "Reset an Interactive Report (IR)", there have been numerous customer discussions and blog posts which show how to directly use the gReport JavaScript object to manipulate an Interactive Report.  The problem?  With the massive rewrite to support multiple Interactive Reports in Oracle Application Express 5, gReport no longer exists.  And as Trent astutely points out, gReport isn't documented.  And that's the cautionary tale here - if it's not documented, it's not considered supported or available for use and is subject to change, effectively without notice.  While I appreciate the inventiveness of others to do amazing things in their applications, and share that knowledge with the Oracle APEX community, you must be cautious in what you adopt.
  2. In the rewrite of Interactive Reports, the IR component was completely revamped from top to bottom.  The markup used for IRs in APEX 5 is dramatically improved:  less tables, much smaller and more efficient markup, better accessibility, etc.  However, if you've also followed this blog post from Shakeeb Rahman (@shakeeb) from 2010, and directly overrode the CSS classes used in Interactive Reports, that will no longer work in IRs in APEX 5.  Your custom styling by using these classes will not have any effect.
  3. As the Oracle Application Express 5 Beta documentation enumerates, there is a modest list of deprecated features and a very small list of features which are no longer supported.  "Deprecated" means "will still work in APEX 5, but will go away in a future release of APEX, most likely the next major release of APEX".  In some cases, like the deprecated page attributes for example, if you have existing applications that use these attributes, they will still function as in earlier releases of APEX, but you won't have the ability to set it for new pages.  Personally, I'm most eager to get rid of all uses of APEX_PLSQL_JOB - customers should use SYS.DBMS_SCHEDULER - it's far richer in functionality.
Please understand that we have carefully considered all of these decisions - even labored for days, in some cases.  And while some of these changes could be disruptive for existing customers, especially if you've used something that is internal and not documented, we would rather have the APEX Community be made aware of these changes up front, rather than be silent about it and hope for the best.

Annonce : Remise du prix de la meilleure thèse

Jean-Philippe Pinte - Tue, 2015-02-10 03:12
Oracle France a remis à M. Gérald Patterson (ISEP / 2ième promotion du Master Cloud Computing) le trophée Oracle pour sa thèse intitulée Improving Cloud Computing availability with Openstack Enhanced Performance.

Ce fut également l'occasion de présenter le aux élèves de l'ISEP.

Exception from executeScript in Alfresco Share

Yann Neuhaus - Tue, 2015-02-10 03:00

I didn't have the opportunity to post a new entry about Alfresco in this blog for a long time now, so I will fix this! In this blog entry, I will talk about a bug I encountered a few months ago. I resolved it but I, so far, not had the time to share my knowledge with you.

I. Description of the issue

This bug appears no matter what the version of Alfresco is used, regardless of the components that are installed, aso... So what is this bug? In fact, this bug isn't blocking anything. Actually it has no impact on the daily work, however, it fills up the Alfresco log files very quickly which can be problematic if you are an administrator searching for information in these log files! Indeed, each time a user accesses a page of Alfresco, between 10 and 50 Java Exceptions are generated (always the same), this create gigabytes log files in minutes/hours. Here is the exception I'm talking about:

Jul 08, 2014 10:42:16 AM org.apache.catalina.startup.Catalina start
INFO: Server startup in 95898 ms
2013-07-08 10:45:02,300 INFO [] [http-apr-8080-exec-1] Successfully retrieved license information from Alfresco.
2013-07-08 10:45:02,417 ERROR [extensions.webscripts.AbstractRuntime] [http-apr-8080-exec-3] Exception from executeScript - redirecting to status template error: 06080001 Unknown method specified to remote store API: has
  org.springframework.extensions.webscripts.WebScriptException: 06080001 Unknown method specified to remote store API: has
  at org.alfresco.repo.web.scripts.bean.BaseRemoteStore.execute(
  at org.alfresco.repo.web.scripts.RepositoryContainer$3.execute(
  at org.alfresco.repo.transaction.RetryingTransactionHelper.doInTransaction(
  at org.alfresco.repo.web.scripts.RepositoryContainer.transactionedExecute(
  at org.alfresco.repo.web.scripts.RepositoryContainer.transactionedExecuteAs(
  at org.alfresco.repo.web.scripts.RepositoryContainer.executeScript(
  at org.springframework.extensions.webscripts.AbstractRuntime.executeScript(
  at org.springframework.extensions.webscripts.AbstractRuntime.executeScript(
  at org.springframework.extensions.webscripts.servlet.WebScriptServlet.service(
  at javax.servlet.http.HttpServlet.service(
  at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(
  at org.apache.catalina.core.ApplicationFilterChain.doFilter(
  at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(
  at org.apache.catalina.core.ApplicationFilterChain.doFilter(
  at org.apache.catalina.core.StandardWrapperValve.invoke(
  at org.apache.catalina.core.StandardContextValve.invoke(
  at org.apache.catalina.authenticator.AuthenticatorBase.invoke(
  at org.apache.catalina.core.StandardHostValve.invoke(
  at org.apache.catalina.valves.ErrorReportValve.invoke(
  at org.apache.catalina.valves.AccessLogValve.invoke(
  at org.apache.catalina.core.StandardEngineValve.invoke(
  at org.apache.catalina.connector.CoyoteAdapter.service(
  at org.apache.coyote.http11.AbstractHttp11Processor.process(
  at org.apache.coyote.AbstractProtocol$AbstractConnectionHandler.process(
  at java.util.concurrent.ThreadPoolExecutor.runWorker(
  at java.util.concurrent.ThreadPoolExecutor$


The first time I encountered this exception, it was on an Alfresco v4.x installation, up and running for some years with a lot of extensions/customizations (mostly .AMP files). If you need more information about AMPs, it means Alfresco Module Package, it's the better way to extend Alfresco. Please take a look at some of my old blogs to find information about how to create this kind of stuff!

I had this exception on more than one Alfresco server, and because of that, I firstly thought that this exception came from an AMP... Therefore, I went through all these extensions but despite hours of research, I found nothing.

II. How to replicate the issue

I tried to replicate this issue with a fresh installation of Alfresco, same version, same extensions, aso... But I haven't been able to do so, at the first place. Finally, one day, I found out that there were something strange with the Alfresco servers on which the Java Exceptions appeared: the "Sites" folder weren't there. Indeed, after the installation of a new Alfresco server, the deletion of the default site ("Sample: Web Site Design Project") and the deletion of the "Sites" folder from the Repository browser, the exception appeared magically in the Alfresco log files... Now that we know from where this issue comes from, it's quite easy to replicate it:

  1. Install a new Alfresco server with the same version from the bundle executable/binaries (quicker)
  2. Start the Alfresco server and open the Alfresco Share UI (http://localhost:8080/share) using the admin account
  3. Navigate to the Sites finder (http://localhost:8080/share/page/site-finder)
  4. Click on "Search" to display all existing sites (only the default one is present: "Sample: Web Site Design Project")
  5. Click on "Delete" to delete the swsdp site
  6. Navigate to the Repository (http://localhost:8080/share/page/repository)
  7. Remove the "Sites" folder on the Repository page (/Company Home/Sites)
  8. Refresh the page and take a look at your logs

After doing that, you should be able to see a lot of exceptions like the one describe above. Issue replicated!

III. How to solve the issue

Being able to replicate an issue is good, but knowing how to solve it is better!

If the "Sites" folder has been deleted in the first place, it was because the Alfresco Sites weren't used at all. Therefore, the simplest solution to resolve this issue was to get the "Sites" folder back. But it's not that easy because this folder has a particular type, some particular attributes, aso... You can't just create a new folder, rename it "Sites" and hope that it will work ;). Starting from here, what you can do to solve this issue is:

  1. Restore the "Sites" folder using a backup
  2. Replicate the "Sites" folder from another Alfresco server

If you don't have a way to restore the "Sites" folder like it was my case (after some months, no backup left), here is what you can do to fix this issue:

Let's say that the Alfresco server, where the "Sites" folder doesn't exist anymore, is named "A". Please take a look at the end of this blog entry for some screenshots that may help you.

  1. Install a new Alfresco server with the same version as "A" from the bundle executable/binaries. This can be on your local machine. Let's name this Alfresco server "B"
  2. Start the Alfresco server "B" and open the Alfresco Share UI (http://localhost:8080/share) using the admin account
  3. Navigate to the Sites finder (http://localhost:8080/share/page/site-finder)
  4. Click on "Search" to display all existing sites (only the default one is present: "Sample: Web Site Design Project")
  5. Click on "Delete" to delete the swsdp site
  6. Navigate to the Repository (http://localhost:8080/share/page/repository) (DON'T delete the "Sites" folder)
  7. Configure a replication target on "B" to point to "A" (take a look at the Alfresco doc:
  8. Enable the replication:
    1. Add the Alfresco Share url and the RepositoryId of "A" into the share-config-custom.xml file of "B" (take a look at the Alfresco doc:
    2. Add the "replication.enabled=true" into the file of "B" (take a look at the Alfresco doc:
    3. Restart "B" for the changes to be taken into account by Alfresco
  9. Configure a replication job on "B" to replicate the "Sites" folder from "B" to "A" (http://localhost:8080/share/page/console/admin-console/replication-jobs)
  10. Run the replication job on "B"

Configure the Replication Target on B (step 7 - create a folder named "TransfertToA" and edit its permissions):


Find the Repository ID of A (step 8.1):


Configure the share-config-custom.xml file of B (step 8.1):


Once the replication job has run on "B", the exceptions will disappear from the log files of "A". I didn't go deeper so I don't really know if you can create new sites using this newly imported "Sites" folder but if you removed this folder in the first place, I would guess that you don't really need it ;).

Thank you for reading this post and I hope this will help. If you need more information, don't hesitate to let a little comment below. See you soon for more blogs!


A Sneak Preview of e-Literate TV at ELI

Michael Feldstein - Tue, 2015-02-10 00:58

By Michael FeldsteinMore Posts (1013)

Phil and I will be chatting with Malcolm Brown and Veronica Diaz about our upcoming e-Literate TV series on personalized learning in a featured session at ELI tomorrow. We’ll be previewing short segments of video case studies that we’ve done on an elite New England liberal arts college, an urban community college, and large public university. Audience participation in the discussion is definitely encouraged. It will be tomorrow at 11:45 AM in California C for those of you who are here at the conference, and also webcast for those of you registered for the virtual conference.

We hope to see you there.

The post A Sneak Preview of e-Literate TV at ELI appeared first on e-Literate.

Fronting Oracle Maven Repository with Artifactory

Steve Button - Mon, 2015-02-09 22:44
The JFrog team announced this week the release of Artifactory 3.5.1, which is a minor update that now works with the Oracle Maven Repository.

I spent a little while yesterday having a look at it, working through the configuration of a remote repository and testing it with a maven project to see how it worked.

Once I'd downloaded it and started it up -- much love for simple and obvious bin/*.sh scripts -- it was a very simple two step process:

1. Since we live behind a firewall first add a proxy configuration to point at our proxy server.

2. Add a new remote repository and pointed it at the Oracle Maven Repository, specifying its URL and using my OTN credentials as username and password.

The Artifactory 3.5.1 documentation stated that the Advanced Settings >  Lenient host authentication and Enable cookie management options must be checked when accessing the Oracle Maven Repository.

The Test button is handy to verify the server settings have been entered correctly.

3. Use the Home tab > Client Settings > Maven Settings link to generate and save a settings.xml file that uses the artifactory server.

With the repository running, configured and the settings.xml saved, its then possible to try it out with an existing maven project such as

I also nuked my local repository to force/verify that the dependencies were fetched through the specified Artifactory server.

$ rm -fr ~/.m2/repository/com/oracle
$ mvn -s artifactory-settings.xml test

Viewing the output of the mvn process and the running Artifactory server you can see that maven is downloading dependencies from http://localhost:8081/artifactory and correspondingly Artifactory is downloading the requested artifact from

Once the maven process has completed and all the requested artifacts downloaded, Artifactory will have cached them locally for future use.
Using the Search functionality of the Artifactory Web UI you can search for weblogic artifacts.

Using the Repository Browser functionality of the Artifactory Web UI you can view and navigate around the contents of the remote Oracle Maven Repository.

Nice JFrog > Artifactory team - thanks for the quick support of our repository.

One further thing I'd look at doing is enabling the Configure Passwords Encryption option in the Security settings to encrypt your OTN password, so that it's not stored in cleartext in the etc/artifactory.config.latest.xml file.

Goldengate – start replicat ATSCN or AFTERSCN ?

Michael Dinh - Mon, 2015-02-09 21:36

When using Goldengate to instantiate target database from an Oracle source database, replicat process can be started to concide with extract based the method used for instantiation, e.g. RMAN or datapump.

ATSCN is used to start replicat if RMAN is used to instantiate target.
From Database Backup and Recovery Reference, UNTIL SCN specifies an SCN as an upper limit.
RMAN restore or recover up to but not including the specified SCN.

AFTERSCN is used to start replicat if datapump is used to instantiate target.
The export operation performed is consistent as of FLASHBACK_SCN.

Hope this helps to clear up when to use ATSCN versus AFTERSCN.

Oracle GoldenGate Best Practices: Instantiation from an Oracle Source Database –  Doc ID 1276058.1

JavaScript Stored Procedures and Node.js Applications with Oracle Database 12c

Kuassi Mensah - Mon, 2015-02-09 20:44
JavaScript Stored Procedures and Node.js Applications with Oracle Database 12c                                       Kuassi Mensah                           | @kmensah |
Keywords: JavaScript, Node.js, Java, JVM, Nashorn, Avatar.jsIntroduction                                                            Node.js and server-side JavaScript are hot and trendy; per the latest “RedMonk Programming Languages Rankings[1], JavaScript and Java are the top two programming languages. For most developers building modern Web, mobile, and cloud based applications, the ability to use the same language across all tiers (client, middle, and database) feels like Nirvana but the IT landscape is not a green field; enterprises have invested a lot in Java (or other platforms for that matter) therefore, the integration of JavaScript with it becomes imperative. WebSockets and RESTful services enable loose integration however, the advent of JavaScript engines on the JVM (Rhino, Nashorn, DynJS), and Node.js APIs on the JVM (Avatar.js, Nodyn, Trireme), make possible and very tempting to co-locate Java and Node applications on the same JVM.
This paper describes the steps for running JavaScript stored procedures[2]directly on the embedded JVM in Oracle database 12c and the steps for running Node.js applications on the JVM against Orace database 12c, using Avatar.js, JDBC and UCP.          
JavaScript and the Evolution of Web Applications Architecture                                   At the beginning, once upon a time, long time ago, JavaScript was a browser-only thing while business logic, back-end services and even presentations where handled/produced in middle-tiers using Java or other platforms and frameworks. Then JavaScript engines (Google’s V8, Rhino) leave the browsers and gave birth to server-side JavaScript frameworks and Node.js.Node Programming ModelNode.js and similar frameworks bring ease of development rapid prototyping, event-driven, and non-blocking programming model[3]to JavaScript. This model is praised for its scalability and good enough performance however, unlike Java, Node lacks standardization in many areas such as database access i.e., JDBC equivalent, and may lead, without discipline, to the so called “callback hell[4]”.
Nonetheless, Node is popular and has a vibrant community and a large set of frameworks[5].Node Impact on Web Applications ArchitectureWith the advent of Node, REST and Web Sockets, the architecture of Web applications has evolved into (i) plain JavaScript on browsers (mobiles, tablets, and desktops); (ii) server-side JavaScript modules (i.e., Node.js, ORM frameworks) interacting with Java business logic and databases.The new proposal for Web applications architecture is the integration of Node.js and Java on the JVM.  Let’s discuss the enabling technologies: JavaScript engine on the JVM and Node API on the JVM and describe typical use cases with Oracle database 12c.  JavaScript on the JVMWhy implement a JavaScript engine and run JavaScript on the JVM? For starters, i highly recommend Mark Swartz ‘s Steve Yegge’s posts. In summary, the JVM brings (i) portability; (ii) manageability; (iii) Java tools; (iv) Java libraries/technologies such as JDBC, Hadoop; and (v) the preservation of investments in Java. 
There are several implementations/projects of Java based JavaScript engines including Rhino, DynJS and Nashorn.Rhino
First JavaScripe engine entirery written in Java; started at NetScape in 1997 then, became an open-source Mozilla project[6]. Was for quite some time the default JavaScript engine in Java SE, now  replaced by Nashorn in Java SE 8. DynJS
DynJS is another open-source JavaScript engine for the JVM. Here is the project homepage Nashorn
Introduced in Java 7 but “production” in Java 8[7], the goal of project Nashorn (JEP 174), is to enhance the performance and security of the Rhino JavaScript engine on the JVM. It integrates with javax.script API (JSR 223) and allows seamless interaction between Java and JavaScript (i.e., invoking Nashorn from Java and invoking Java from Nashorn).

To illustrate the reach of Nashorn on the JVM and the interaction between Java and JavaScript, let’s run some JavaScript directly on the database-embedded JVM in Oracle database 12c. JavaScript Stored Procedures with Oracle database 12c Using Nashorn
Why would anyone run JavaScript in the database? For the same reasons you’d run Java in Oracle database. Then you might ask why run Java in the database, in the first place? As discussed in my book[8], the primary motivations are: (i) reuse skills and code, i.e., which programming languages are your new hire knowledgeable of or willing to learn; (ii) avoid data shipping[9] i.e., in-place processing of billions of data/documents; (iii) combine SQL with foreign libraries to achieve new database capability thereby extending SQL and the reach of the RDBMS, e.g., Web Services callout, in-database container for Hadoop[10]. Some developers/architects prefer a tight separation between the RDBMS and applications therefore, no programming language in the database[11]but there are many pragmatic developers/architects who run code near data, whenever it is more efficient than shipping data to external infrastructure.
Co-locating functions with data on the same compute engine is shared by many programming models such as Hadoop. With the surge and prevalence of Cloud computing, RESTful service based architecture is the new norm. Data-bound services can be secured and protected by the REST infrastructure, running outside the RDBMS. Typical use case: a JavaScript stored procedures service would process millions/billions of JSON documents in the Oracle database and would return the result sets to the service invoker.
To conclude, running Java, JRuby, Python, JavaScript, Scala, or other programming language on the JVM in the database is a sound architectural choice. The best practices consist in: (i) partitioning applications into data-bound and compute-bound modules or services; (ii) data-bound services are good candidates for running in the database; (iii) understand DEFINER’s vs INVOKER’s right[12]and grant only the necessary privilege and/or permission. 

The Steps
The following steps allow implementing JavaScipt stored procedure  running in Oracle database; these steps represent an enhancement from the ones presented at JavaOne and OOW 2014 -- which consisted in reading the JavaScript from the database file system; such approach required granting extra privileges to the database schema for reading from RDBMS file system something not recommended from security perspective. Here is a safer approach:

1.      Nashorn is part of Java 8 but early editions can be built for Java 7; the embedded JavaVM in Oracle database 12c supports Java 6 (the default) or Java 7. For this proof of concept, install Oracle database 12c with Java SE 7 [13]2.      Build a standard Nashorn.jar[14]; (ii) modify the Shell code to interpret the given script name as an OJVM resource; this consists mainly in invoking getResourceAsStream()on the current thread's context class loader ; (iii) rebuild Nashorn.jar with the modified Shell 3.  Load the modified Nashorn jar into an Oracle database shema e.g., HR
 loadjava -v -r -u hr/ nashorn.jar4.      Create a new dbms_javascript  package for invoking Nashorn’s Shell with a script name as parameter
create or replace package dbms_javascript as
  procedure run(script varchar2);
create or replace package body dbms_javascript as
  procedure run(script varchar2) as
  language java name '[])';
Then call dbms_javascript,run(‘myscript.js’)from SQL which will invoke Nashorn  Shell to execute the previously loaded myscript.js.5.  Create a custom role, we will name it NASHORN, as follows, connected as SYSTEM
SQL> create role nashorn;
SQL> call dbms_java.grant_permission('NASHORN', 'SYS:java.lang.RuntimePermission', 'createClassLoader', '' );
SQL> call dbms_java.grant_permission('NASHORN', 'SYS:java.lang.RuntimePermission', 'getClassLoader', '' );
SQL> call dbms_java.grant_permission('NASHORN', 'SYS:java.util.logging.LoggingPermission', 'control', '' );Best practice: insert those statements in a nash-role.sqlfile and run the script as SYSTEM6.      Grant the NASHORN role created above to the HR schema as follows (connected as SYSTEM):

SQL> grant NASHORN to HR;

7.      Insert the following JavaScript code in a file e.g., database.js stored on your client machine’s (i.e., a machine from which you will invoke loadjava as explained in the next step).
This script illustrates using JavaScript and Java as it uses the server-side JDBC driver to execute a PreparedStatement to retrieve the first and last names from the EMPLOYEES table.var Driver =;
var oracleDriver = new Driver();
var url = "jdbc:default:connection:";   // server-side JDBC driver
var query ="SELECT first_name, last_name from employees";
// Establish a JDBC connection
var connection = oracleDriver.defaultConnection();
// Prepare statement
var preparedStatement = connection.prepareStatement(query);
// execute Query
var resultSet = preparedStatement.executeQuery();
// display results
     while( {
     print(resultSet.getString(1) + "== " + resultSet.getString(2) + " " );
// cleanup

8.      Load database.js in the database as a Java resource (not a vanilla class)
loadjava –v –r –u hr/ database.js9.      To run the loaded scriptsqlplus hr/
SQL>set serveroutput on
SQL>call dbms_java.set_output(80000)
SQL>call‘database.js’);The Nashorn Shell reads ‘database.js’ script stored as Java Resource from internal table; the JavaScript in its turn invokes JDBC to execute a PreparedStatement and the result set is displayed on the console. The message “ORA=29515: exit called from Java code with status 0” is due to the invocation of java.lang.Runtime.exitInternal; and status 0 means normal exit (i.e., no error). The fix is to remove that call from Nashorn. Node.js on the JVMAs discussed earlier, Node.js is becoming the man-in-the-middle between Web applications front ends and back-end legacy components and since companies have invested a lot in Java, it is highly desirable to co-locate Node.js and Java components on the same JVM for better integration thereby eliminating the communication overhead. There are several projects re-implementing Node.js APIs on the JVM including: Avatar.js, Nodyn, and Trireme. This paper will only discuss Oracle’s Avatar.js.Project Avatar.js[15]The goal of project Avatar.js is to furnish “Node.js on the JVM”; in other words, an implementation of Node.js APIs, which runs on top of Nashorn and enables the co-location of Node.js programs and Java components. It has been outsourced by Oracle under GPL license[16]. Many Node frameworks and/or applications have been certified to run unchanged or slightly patched, on Avatar.js.
There are binary distributions for Oracle Enterprise Linux, Windows and MacOS (64-bits). These builds can be downloaded from Search for avatar-js.jar and platform specific libavatar-js libraries (.dll, .so, dylib). Get the latest and rename the jar and the specific native libary accordingly. For example: on  Linux, rename the libary to; on Windows, rename the dll to avatar-js.dll and add its location to your PATH (or use -Djava.library.path=).
RDBMSes in general and Oracle database in particular remain the most popular persistence engines and there are RDBMS specific Node drivers[17]as well as ORMs frameworks. However, as we will demonstrate in the following section, with Avatar.js, we can simply reuse existing Java APIs including JDBC and UCP for database access.
Node Programming with Oracle Database using Avatar.js, JDBC and UCP The goal of this proof of concept is to illustrate the co-location of a Node.js application, the Avatar.js library, the Oracle JDBC driver and the Oracle Universal Connection Pool (UCP) on the same Java 8 VM.The sample application consists in a Node.js application which performs the following actions: (i) Request a JDBC-Thin connection from the Java pool (UCP)(ii)Create a PreparedStatement object for “SELECT FIRST_NAME, LAST_NAME FROM EMPLOYEES”
(iii)Execute the statement and return the ResultSet in a callback
(iv)Retrieve the rows and display in browser on port 4000(v) Perform all steps above in a non-blocking fashion – this is Node.js’s raison d’être. The demo also uses Apache ab load generator to simulate concurrent users running the same application in the same/single JVM instance.For the Node application to scale in the absence of asynchronous JDBC APIs, we need to turn synchronous calls into non-blocking ones and retrieve the result set via callback.Turning Synchronous JDBC Calls into Non-Blocking CallsWe will use the following wrapper functions to turn any JDBC call into a non-blocking call i.e., put the JDBC call into a thread pool and free up the Node event loop thread.var makeExecutecallback = function(userCallback) {
 return function(name, args){
      userCallback(undefined, args[1]);
} function submit(task, callback, msg) {
    var handle = evtloop.acquire();
    try {    var ret = task();
      EventType(msg, callback, null, ret)); {catch{}    evtloop.submit(r);
Let’s apply these wrapper functions to executeQuery JDBC call, to illustrate the conceptexports.connect = function(userCallback) {..} // JDBC and UCP settingsStatement.prototype.executeQuery = function(query, userCallback) {
         var statement = this._statement;
          var task = function() {
          return statement.executeQuery(query);
     submit(task, makeExecutecallback(userCallback), "jdbc.executeQuery");
} Similarly the same technique will be applied to other JDBC statement APIs.Connection.prototype.getConnection = function() {…}Connection.prototype.createStatement = function() {..}Connection.prototype.prepareCall = function(storedprocedure) {..}Statement.prototype.executeUpdate = function(query, userCallback) {..}Returning Query ResultSet through a CallbackThe application code fragment hereafter shows how: for every HTTP request: (i) a connection is requested, (ii) the PreparedStatement is executed, and (iii) the result set printed on port 4000.   ...   var ConnProvider = require('./connprovider').ConnProvider;
var connProvider = new ConnProvider(function(err, connection){.. });

var server = http.createServer(function(request, response) {
  connProvider.prepStat(function(resultset) {
                while ( {
                   response.write(resultset.getString(1) + " --" + resultset.getString(2));
server.listen(4000, '');
Using Apache AB, we were able to scale to hundreds of simultaneous invocations of the Node application. Each instance grabs a Java connection from The Universal Connection Pool (UCP), executes the SQL statements through JDBC then return the result set via a Callbak on port 4000.ConclusionsThrough this paper, i discussed the rise of JavaScript for server-side programming and how Java is supporting such evolution; then – something we set out to demonstrate – furnished step by step details for implementing and running JavaScript stored procedures in Oracle database 12c using Nashorn as well as running Node.js applications using Avata.js, Oracle JDBC, UCP against Oracle database 12c.As server-side JavaScript (typified by Node.js) gains in popularity it’ll have to integrate with existing components (COBOL is still alive!!). Developers, architects will have to look into co-locating JavaScript with Java, across middle and database tiers.

[1] [2] I’ll discuss the rationale for running programming languages in the database, later in this paper. [3] Request for I/O and resource intensive components run in separate process then invoke a Callback in the main/single Node  thread, when done. [4] [5] Search the web for “Node.js frameworks[6] [7] Performance being one of the most important aspect [8] [9] Rule of thumb: when processing more than ~20-25% of target data, do it in-place, where data resides (i.e., function shipping). [10] In-database Container for Hadoop is not available, as of this writing. [11] Other than database’s specific procedural language, e.g.,  Oracle’s PL/SQL [12] I discuss this in chapter 2 of my book; see also Oracle database docs. [13] See Multiple JDK Support in [14] Oracle does not furnish a public download of Nashorn.jar for Java 7; search “Nashorn.jar for Java 7”. [15] [16] [17] The upcoming Oracle Node.js driver was presented at OOW 2014. 

EMEA Exadata, Manageability & Hardware Partner Forum

Oracle EMEA Exadata, Manageability, Servers & Storage Partner Community Forum ...

We share our skills to maximize your revenue!
Categories: DBA Blogs

Oracle Maven Repository - Viewing Contents in Eclipse

Steve Button - Mon, 2015-02-09 16:18
With the Oracle Maven Repository now accessible one way to have explore its contents is to use the Maven Repositories viewer feature available in most development tools. I've seen the repository contents displayed easily in NetBeans so I decided to take a look at what it looks like in Eclipse as well.

I had to make a few minor setting changes to get it to work so decided to document them here.  If you've gotten it to work with less setting changes, let me know!

As initial setup, I configured my local maven environment to support access to the Oracle Maven Repository.  This is documented here  I also installed maven-3.2.5 that includes the updated Wagon module that supports authentication.

Next I downloaded and used the new network installer that the Oracle Eclipse team has published on OTN to install the latest version of Oracle Enterprise Pack for Eclipse.

This network installer lets developers select the version of Eclipse to install and the set of Oracle extensions --  Weblogic, GlassFish and other stuff -- to add in to it.

 Once Eclipse is installed, you can add the Maven Repository viewer by selecting   Window > Show View > Other > Maven Repositories from the Eclipse toolbar.

I also added a Console > Maven viewer to see what was happening under the covers and arranged them so they were visible at the same time:

With the Maven views ready to go, expand the Global Repositories node. This will show Maven Central (any other repositories you may have configured) and the Oracle Maven Repository if you have configured it correctly in the settings.xml file.

The initial state of the Oracle Maven Repository doesn't show any contents indicating that its index hasn't been downloaded to display.

Right mouse clicking on it and selecting the Rebuild Index option causes an error to be shown in the console output indicating that the index could not be accessed.

To get it to work, I made the following changes to my environment.  
Configure Eclipse to Use Maven 3.2.5Using the Eclipse > Preferences > Maven > Installation dialog, configure Eclipse to use Maven 3.2.5.  This is preferred version of Maven to use to access the Oracle Maven Repository since it automatically includes the necessary version of the Wagon HTTP module that supports the required authentication configuration and request flow.

Configure Proxy Settings in Maven Settings File ** If you don't need a proxy to access the Internet then step won't be needed **
If you sit behind a firewall and need to use a proxy server to access public repositories then you need to configure a proxy setting inside the maven settings file.

Interestingly for command line maven use and NetBeans a single proxy configuration in settings.xml was enough to allow the Oracle Maven Repository to be successfully accesses and its index and artifacts used.

However with Eclipse, this setting alone didn't allow the Oracle Maven Repository to be accessed.  Looking at the repository URL for the Oracle Maven Repository you can see ity's HTTPS based -- and it appears for Eclipse that a specific HTTPS based proxy setting is required for Eclipse to access HTTPS based repositories.

Rebuild Index SuccessWith the settings in place, the Rebuild Index operation succeeds and the contents of the Oracle Maven Repository are displayed in the repository viewer.

Have your say ...

Tim Dexter - Mon, 2015-02-09 15:25

Another messaging exchange last week with Leslie ...

OK, so we practised it a bit after our first convo and things got a little cheesy but hopefully you get the message.

Hit this link and you too can give some constructive feedback on the Oracle doc for BI (not just BIP.) I took the survey; its only eight questions or more if you want to share more of your input. Please take a couple of minutes to help us shape the documentation of future. 

Categories: BI & Warehousing

Webcast: Delivering Next-Gen Digital Experiences - 2/12

WebCenter Team - Mon, 2015-02-09 15:06
Oracle Corporation Digital Strategies For Customer Engagement Growth Automating Marketing & Customer Engagement

Becoming a digital business is imperative for organizations to deliver the next wave of revenue growth, service excellence and business efficiency. And the stakes are high -- 94% of customers discontinue communications because of irrelevant messages and experiences.

Join this webcast for an in-depth look at technologies that enable IT leaders to connect digital customer experiences to business outcomes. Learn how to:
  • Deliver omni-channel experiences that are seamless, tailored and innovative across Paid, Owned and Earned media
  • Convert unknown audiences to known and involved customers
  • Extend reach and orchestrate engagement across all channels and devices
  • Move Marketing from silo’d technologies to a single Digital Experience Platform that also connects Marketing to the entire organization
Register now for this webcast.

Red Button Top Register Now Red Button Bottom Live Webcast Calendar February 12, 2015
10 a.m. PT / 1 p.m. ET Featured Speaker:

Chris Preston Chris Preston,
Sr. Director
Customer Strategies
Oracle Hardware and Software Engineered to Work Together Copyright © 2015, Oracle Corporation and/or its affiliates.
All rights reserved.
Contact Us | Legal Notices | Privacy

Video Tutorial: XPLAN_ASH Active Session History - Part 3

Randolf Geist - Mon, 2015-02-09 15:04
The next part of the video tutorial explaining the XPLAN_ASH Active Session History functionality continuing the actual walk-through of the script output.

More parts to follow.

Long parse time with interval partitioning and subpartitions

Bobby Durrett's DBA Blog - Mon, 2015-02-09 13:52

In an earlier post I described how some queries with long parse times were causing long library cache lock waits.  Friday I applied a patch from Oracle that resolved the long parse times.  Here are the conditions which may relate to this bug:

  1. Interval partitioned table
  2. Partitioned by range
  3. Sub-partitioned by list
  4. Open ended range on partitioning column in where clause
  5. Tens of thousands of sub-partitions

Prior to applying the patch I did an explain plan on two versions of the problematic query.  One version specified an open-ended range on the partitioning column and the other closed off the range.

Slow version:

T392658.CLNDR_DT >= TO_DATE('2014-11-17' , 'YYYY-MM-DD') and 

Elapsed: 00:00:46.20

Fast version:

T392658.CLNDR_DT >= TO_DATE('2014-11-17' , 'YYYY-MM-DD') and
T392658.CLNDR_DT <= TO_DATE('2014-11-26' , 'YYYY-MM-DD') and

Elapsed: 00:00:00.09

The queries are the same except for the extra date condition which closes off the date range.  Note that the explain plan took 46 seconds with the open-ended range and less than a tenth of a second with the closed off range.

With the patch the slow version is just as fast as the fast one.

This is bug 20061582 in Oracle’s system.

From my conversations with Oracle it sounds like in certain cases the optimizer is reviewing information for many or all of the sub-partitions and since we have 20,000 for this table in production it can take minutes to parse.  I also messed with the dates on the open-ended condition and found that if I made the date early enough the parse time issue went away.  So, it seems that there is some set of conditions, which I don’t know how to easily reproduce, which lead the optimizer to look at sub-partition information and slows parse time.

This is on on an Exadata system.  I was able to reproduce the long parse times on non-Exadata Linux and systems so it does not occur only on Exadata.

This issue surprised me because I thought that the optimizer would not look at partition or sub-partition statistics in a query that has a range condition which spanned more than one partition.  In the past I have always seen the global or table level stats used in these situations.  But, now I know that there are some cases where the optimizer will dig into the sub-partition stats even though the query covers more than one partition.

Looking at the plan of my slow parsing query before and after the patch I get a clue that the optimizer is looking at sub-partition stats:

Partition range without the patch:


Partition range with the patch:

     1 |   136

Evidently KEY(SQ) relates to some sort of filtering of the sub-partitions which cause the long parse time.  The manuals describe KEY(SQ) in a section titled “Dynamic Pruning with Subqueries” so maybe the problem queries have some sub-query that the optimizer was using to choose which sub-partitions that the query needed.

If you have an interval partitioned table with tens of thousands of sub-partitions and parse times in the minutes with open-ended ranges specified on the partitioning column your system could be hitting this same issue.

– Bobby

Categories: DBA Blogs

MAF 2.1 Alta Mobile UI and Oracle Mobile Suite

Andrejus Baranovski - Mon, 2015-02-09 12:19
This post is about Alta Mobile UI and MAF 2.1. I was using Oracle Work Better mobile application as a reference, along with Alta Mobile UI Design Guidelines. This is based on my previous posts about Oracle Mobile Suite and MAF 2.1 integration, read more here - Oracle Mobile Suite - Web Service Performance Optimisation with Result Caching. You could also use Alta UI 12c with regular ADF Web applications, check more here - Custom ADF Application with New ADF 12c Alta UI.

Sample application - (contains ADF BC SOAP Web Service, Oracle Mobile Suite SOAP to REST transformation and MAF 2.1 application with Alta Mobile UI) implements a dashboard. Component to visualise amounts and values comes out of the box with MAF (iPad view):

User could open a springboard and select Employees option in the sample application:

Here we can search for employees by first name/last name and bring results as a list:

Optionally user could switch to cards view, to display the same results in different way:

You could click on employee icon - this will load details screen and bring additional info:

Steven King is a manager, he manages Executive department and there are two team members reporting to him. Compensation graph can be displayed for team members:

A list of team members is displayed in the third tab:

There is MAF 2.1 Deck component in Employees search page, it allows to switch between a list and cards view:

Employee details page is rendered with a help of MAF 2.1 Deck component as well, along with Select Button component to display tab UI and navigate between Detail, Compensation and Team sections:

You could check how these pages are implemented in the sample app provided above.

Flat World and CBE: Self-paced does not imply isolation

Michael Feldstein - Mon, 2015-02-09 07:48

By Phil HillMore Posts (287)

As competency-based education (CBE) becomes more and more important to US higher education, it would be worth exploring the learning platforms in use. While there are cases of institutions using their traditional LMS to support a CBE program, there is a new market developing specifically around learning platforms that are designed specifically for self-paced, fully-online, competency-framework based approaches.

Several weeks ago Flat World announced their latest round of funding, raising $5 million of debt financing, raising their total to $40.7 million. The company started out by offering full e-textbooks (and was previously named FlatWorld Knowledge), developing 110 titles that included 25 of the 50 most-used lecture courses. The e-textbook market was not working out, however, and the company pivoted to competency-based education around the time that Chris Etesse became CEO two years ago. Now the company is developing a combined CBE learning platform with integrated course content – much of it repurposing the pre-existing e-textbook materials. Their first academic partner for CBE is Brandman University, a non-traditional part of the Chapman University system and is currently one of the CBEN network.

One central tenet of the Flat World approach is based on their history and pivot – a tight integration of content and platform. As Etesse describes it, content is a 1st-class citizen in their system whereas other loosely-coupled approaches that do not tie content and platform together can be difficult to navigate and collect learning analytics. In other words, this intentionally is a walled-garden approach. For Brandman, approximately 70% of the content comes from the pre-existing FlatWorld texts, 25% comes from various OER sources, and about 5% has been custom-designed for Brandman.

In other words, this is very much a walled garden by design. While there is support for outside content, I believe this integration must be done by Flat World designers.

As was the case for the description of the Helix CBE-based learning platform, my interest here is not merely to review one company’s products, but rather to illustrate aspects of the growing CBE movement using the demo.

CBE programs by their very nature tend to be self-paced. One criticism or line of questions I’m seeing more often deals with the nature of self-paced learning itself. Are students just plugging through mindless e-text and multiple-choice assessments in isolation? What Flat World illustrates – as with other major CBE learning platforms – is that self-paced does not imply isolation, either from a student-teacher or from a student-student perspective. New approaches that are different than simple discussion forums are required, however.

FlatWorld shows several learning activities:

  • Reading text and viewing multi-media content adaptively presented based on a pretest and progress against competencies;
  • Taking formative assessments primarily through multiple-choice quizzes;
  • Interacting with students and with faculty;
  • Working through project-based assignments;
  • Taking summative assessments through proctored, webcam-streaming approach.

The activities and assessments do not have to be students working in isolation and using multiple-choice. For example, the project based work can be included and assignments can include submission of written reports or based on short-form prompts. As can be seen below, the assessments can be based on submitted written work which faculty grade and use for feedback.


For communication with others, students are tracked in how active they are in communicating with faculty and even with other students (called ‘social’), as seen below.


One challenge of a self-paced program such as CBE approaches is figuring out how to encourage students to interact with others. There is not a simple cohort to work with – the interaction instead will often be based on content. Who else is working through the same content in roughly the same time period.

FlatWorld uses an approach that is very similar to Stack Overflow, where students can ask and answer questions over time, and the answers are voted up or down to allow the best answers to rise to the top. The stack overflow is moderated by faculty at Brandman. This not only allows students working on the same competencies at roughly the same time, but it even allows interaction with students on similar competencies separated in time.




There certainly is a tendency in many CBE programs to stick to multiple-choice assignments and quizzes and to avoid much social interaction. That method is a whole lot easier to design, and with several hundred of new programs under development, I think the overall quality can be quite low in many programs, particularly those looking for a quick-win CBE introduction, essentially trying to jump on the bandwagon. You can see the tendency towards multiple-choice in the FlatWorld system as well.

But self-paced does not imply isolation, and the Flat World implementation of the Brandman University program shows how CBE can support project-based work, written assignments and assessments, and interaction between students and faculty as well as between multiple students.

The post Flat World and CBE: Self-paced does not imply isolation appeared first on e-Literate.

Will Oracle Linux experts need a rundown of Docker?

Chris Foot - Mon, 2015-02-09 01:58

Container technology solution Docker was the talk of the industry last year, as the open source project partnered with IBM, Microsoft and a list of other tech companies in order to reduce server energy usage and drastically simplify software development.

Now, Oracle is hopping on the Docker bandwagon, making its Oracle Linux distribution available in a Docker container on the Docker Hub, according to InformationWeek's Charles Babcock. The goal is to make Oracle Linux more appealing to developers looking to transport apps across different environments. 

Ready for download 
This isn't the first Linux distro to be offered through Docker Hub. Babcock noted that Canonical's Ubuntu had integrated Docker support before Oracle (although Ubuntu is an operating system designed for desktops). In addition, server-based Linux OS distros SUSE, CentOS and Red Hat Enterprise Linux are now available via Docker Hub. 

This poses an interesting question as to how Oracle Linux support groups will integrate Docker's technology into client infrastructures. Think of the impact this development will have on Web-based businesses. Data Center Knowledge noted that more than 33 percent of the globe's websites operate on servers running off of Linux. Therefore, websites running on Oracle Linux may grow leaner, enabling developers to further test the limits of how many features and back-end functions they can add to websites. 

Not the only Oracle product 
Oracle isn't necessarily late to the game as far as Docker integration is concerned. Oracle also owns open source database MySQL, which can be downloaded from Docker Hub. Oracle Senior Vice President of Linux and Virtualization Engineering Wim Coekaertz stated how MySQL's Docker compatibility will impact DBAs who use the database engine. 

"With Oracle Linux and MySQL images available on the Docker Hub Registry, users will be able to quickly create and publish custom Docker containers that layer applications on top of Oracle Linux and MySQL, which is a great time-save for both independent software vendors and IT departments," said Coekaertz, as quoted by DCK.

Coekaertz expanded on the issue, asserting that although Docker promises to simplify virtual machines, he doesn't believe the open source project will deliver "complete isolation" to those who download the Dockerized version of Oracle Linux. 

As far as MySQL is concerned, he believes a Docker-compatible MySQL engine will enhance DevOps staff by allowing engineers to share the container across different teams. This is a role that will help Oracle experts accommodate companies looking to support hybrid environments. 

The post Will Oracle Linux experts need a rundown of Docker? appeared first on Remote DBA Experts.

Never gather WORKLOAD stats on Exadata...

Yann Neuhaus - Mon, 2015-02-09 01:56

For Exadata, oracle has introduced an 'EXADATA' mode which sets a high transfer rate (with IOTFRSPEED as in NOWORKLOAD statistics) and set a MBRC (as in WORKLOAD statistics). Those values are set rather than gathered because all the SmartScan optimization done at storage cell level, which makes the multiblock reads less expensive, is difficult to measure from the database.
Here I will explain what I stated in a previous blog: direct-path reads are not counted as multiblock reads for the MBRC system statistic. And direct-path read should be the main i/o path in Exadata as you probably bought that machine to benefit from SmartScan.

With direct-path reads

On a test database that has no activity, I’m creating a 1000 blocks table. My goal is to gather WORKLOAD system statistics during a simple table full scan on that table, and see how it calculates SREADTIM, MREADTIM and MBRC.

SQL> connect demo/demo

SQL> drop table DEMO;
Table dropped.

SQL> create table DEMO pctfree 99 as select rpad('x',1000,'x') n from dual connect by level <=1000;
Table created.
Then I run a simple select between the calls to ‘start’ and ‘stop’ procedures of the dbms_stats WORKLOAD system stats gathering.
SQL> exec dbms_stats.gather_system_stats('start');
PL/SQL procedure successfully completed.

SQL> connect demo/demo

SQL> select count(*) from DEMO;


I check the physical read statistics (this is why have reconnected my session so that I can query v$mystat without doing the delta)
SQL> select name,value from v$mystat join v$statname using(statistic#) where (name like 'phy%' or name like 'cell%') and value>0;

NAME                                                              VALUE
------------------------------------------------------------ ----------
physical read total IO requests                                      22
physical read total multi block requests                              7
physical read total bytes                                       8306688
cell physical IO interconnect bytes                             8306688
physical reads                                                     1000
physical reads direct                                              1000
physical read IO requests                                            15
physical read bytes                                             8192000
cell scans                                                            1
cell blocks processed by cache layer                               1000
cell blocks processed by txn layer                                 1000
cell blocks processed by data layer                                1000
cell physical IO bytes eligible for predicate offload           8192000
cell physical IO interconnect bytes returned by smart scan       130760
cell IO uncompressed bytes                                      8192000
I’ve read 1000 blocks in 15 i/o calls so I'm sure it is multiblock reads. All of them (1000 x 8k) was eligible for SmartScan and those 1000 blocks have been processed by the storage cell.

If you wonder why I have only 7 'physical read total multi block requests' it's because it accounts only the 'full' multiblock reads - not those that are limited by extent boundary. See here for that analysis.

If you wonder why I have only 22 'physical read total IO requests' then I've not the answer. The sql_trace shows only the 15 'direct path read'. And dbms_stats counts only the 'physical read IO requests'. If you have any idea, please comment.

I stop my WORKLOAD statistics gathering:
SQL> exec dbms_stats.gather_system_stats('stop');
PL/SQL procedure successfully completed.
And check the system statistics that have been set:

SQL> select * from sys.aux_stats$;

SNAME           PNAME           PVAL1
--------------- ---------- ----------
SYSSTATS_INFO   FLAGS               1
I have no SREADTIM which is expected as I've done only multiblock reads. I have a MREADTIM. But I don't have the MBRC set.

With conventional (aka buffered) reads

Let's do the same after disabling serial direct-path reads:

SQL> alter session set "_serial_direct_read"=never;
Session altered.
I do the same as before, but now my session stats show only conventional reads:
NAME                                                              VALUE
------------------------------------------------------------ ----------
physical read total IO requests                                      44
physical read total multi block requests                             28
physical read total bytes                                       8192000
cell physical IO interconnect bytes                             8192000
physical reads                                                     1000
physical reads cache                                               1000
physical read IO requests                                            44
physical read bytes                                             8192000
physical reads cache prefetch                                       956

and here are the gathered stats:
SNAME           PNAME           PVAL1
--------------- ---------- ----------
SYSSTATS_INFO   FLAGS               1
SYSSTATS_MAIN   MBRC               23

Now the MBRC is set with the gathered value.

This proves that MBRC is set only for conventional multiblock reads. Direct-path reads are not accounted.

Conclusion If you are on Exadata, you probably want to benefit from SmartScan. Then you probably want the CBO to choose FULL TABLE SCAN which will do direct-path reads for large tables (according that they don't have a lot of updated buffers in SGA). If you gather WORKLOAD statistics they will set MBRC without accounting for those direct-path reads and it will probably be set lower than the average actual multiblock read (which - in direct-path reads - is close the the db_file_multiblock_read - default or set value).
This is the reason why Oracle introduced the EXADATA mode: it sets the MBRC from the db_file_multiblock_read value. They also set the IOTFRSPEED to a high value because gathering MREADTIM will probably get a very low value - lower than SREADTIM - thanks to the SmartScan. And CBO ignores values where MREADTIM is less than SREADTIM.

An alternative to EXADATA mode can be setting those values as NOWORKLOAD statistics and keep the db_file_multiblock_read_count set. You will have the same behavior because CBO uses db_file_multiblock_read_count when it is set and there are no MBRC system stats. But the danger is that if someone resets the db_file_multiblock_read_count (and I often advise to keep defaults) then the CBO will use a value of 8 and that will probably increase the cost of full table scans too much.

All formulas are here with a script that shows what is used by the CBO.

Never say never Well, that blog post title is too extreme because:

@FranckPachot @kevinclosson @fritshoogland @moustafa_dba @JLOracle never say never ;-)

— Christian Antognini (@ChrisAntognini) February 8, 2015 So I should say:
Never gather WORKLOAD stats on Exadata... except if your workload is not an Exadata optimized one.
If you are using Exadata for OLTP, then yes, you can gather WORKLOAD statistics as they probably fit OLTP behaviour. But in any case, always check the gathered stats and see if they are relevant.