BI & Warehousing

Oracle Data Integrator is being used as the ETL tool for BI Apps.  Here are some benefits: 1. Productivity It tasks less efforts for developers to write a ETL map comparing  to other ETL tool. 2. Define ETL flow patterns in KM The reason why now we can take less effort is that the flow […]

I had a question last week from a great customer that is looking to build some custom .Net apps that enable users to jump from the app to the BIP interface with a click. The fly in the ointment, no single sign on (SSO) servers in sight! Day-um!

Some messing with the BIP login and some knowledge of the underlying jsp files that are actually getting called i.e. http://server:port/xmlpserver is actually resolving to http://server:port/xmlpserver/login.jsp. Then some use of the IE/FF Developer tools to find the appropriate names for the user name and password fields on the login page and I came up with:

http://server:port/xmlpserver/login.jsp?id=<<username>>&passwd=<<password>>

for instance

http://localhost:7001/xmlpserver/login.jsp?id=tdexter&passwd=bipdex

It worked! Sometimes, I even amaze myself at my detective (read 'hacking') abilities :0)

A request from Leslie today to help her out on the user docs. In them we state that we support the MSWord organization charts but we do not give any detail.

Use the organization chart functionality in the templates and the chart that is rendered in the output. Figure 4-18 shows an example of an organization chart.

Figure 4-18 Sample Organization Chart

Its been a  while since I have looked at them but we mean just that. You build an org chart with names in the boxes, BIP will render it, simple.

Oh, you wanted it to load the names into the chart dynamically from the dataset? Sorry, no dice, at least not with the MSWord Org Chart object.

However, you can create your own org chart structure using MSShapes and use BIP's ability to fill those shapes with text from your data. Thats documented pretty well and is very easy to do. Taking it to obvious final step; completely data driven org chart structure and text. Thats a bit tougher. It can be done with the shape copy and move commands but its going to take some planning. You need to think about how wide your 'page' is, what to do when you reach the edge and need to continue with the same level in the hierarchy, etc.

To get you started, I have created a sample template and data for the first two scenarios. They will work with all releases of BIP and XMLP. The third will take me a little longer :0)

Oracle EPM 11.1.2.3 is out and there are some great features in it (Planning and Financial Reporting have my favorites which you'll see in a second).  11.1.2.3 is not as impressive as 11.1.2.1 or 11.1.2.2, though.  A lot of the products got new features in between releases like Smart View (which adds new features with every patch set), Strategic Finance, HFM (which got Tax Provisioning in February), and Financial Close Management instead of making those products wait until 11.1.2.3 was officially released.

Below are the features I think are most interesting in 11.1.2.3.  This is by no means an exhaustive list and I didn't include every product though I did hit all the ones with significant market share (and a few others including a shout-out to the long-forgotten "EPM Workspace").  All the new feature documents are on Oracle.com.  If you see anything major I missed, let me know in the comments.


Oracle Essbase

• Essbase didn't change much in 11.1.2.3 and users are never going to see any of the changes.
• There are several new BSO functions and calc script commands like @INTERSECT, @ISMBRUDA, @ISRANGENONEMPTY, @MEMBERAT, @RANGEFIRSTVAL, and @RANGELASTVAL.  My favorite is @CREATEBLOCK which, wait for it, creates a sparse block with all the dense combinations in it set to missing.  Yes, we finally have a specific command to create specific blocks after 20+ years of silly block creation workarounds.
• There's a new calc script SET command called RUNTIMESUBVARS that let's you create substitution variables in the script that are passed to the script at run-time.  You can then pass values to these variables when you call the calc scripts from MaxL or the API.  There's a new optional argument (with runtimesubvars) to MaxL's "execute calculation" command that sends the run-time variables to the calc script as you call it.
• There's a new Essbase.CFG setting called ESTIMATEDHASHSIZE that lets you specify (in millions) how many members should be loaded into memory from the outline.  It's meant to speed up massive dimension builds and outline edits.  There are two other new CFG settings as well: ENABLERTSVLOGGING (which logs the new run-time substitution variables) and UNICODEENABLE (which sets the server to unicode mode, basically).
• Data Mining is gone.  Many people didn't even know it was there, but it's been removed.
• Speaking of gone, there is no Essbase Excel Add-In 11.1.2.3.  Further, earlier versions of the Add-In that you may have (like 11.1.2.2) are not supported against Essbase 11.1.2.3.  It probably will work (the Add-In is wonderfully backwards and forwards compatible) but it's not supported.

Smart View

• There were a lot of improvements to Smart View in 11.1.2.2.3xx which came out about a month ago, so there isn't much new in Smart View 11.1.2.3.  Smart View is the new strategic Microsoft Office add-in for all the Business Analytics (BI and EPM) products, though, so that's why the newest release supported direct access to OBIEE 11.1.1.7.  Hyperion Strategic Finance is also now fully supported in Smart View.
• The only major improvement is actually in Hyperion Planning just exposed through Smart View.  There is a new "Planning Admin" extension that lets administrators update Planning metadata (and a few other things) from Smart View.  I'll say more about this under Planning.

Hyperion Planning

• Essbase ASO databases can now be created as Planning Plan Types.  These databases will allow writeback provided it's at level-0.  This is stunning.  These ASO cubes can be displayed on Planning composite forms with Essbase BSO cubes just like current BSO cubes in multiple plan types can be combined on a single composite form.  For some reason, this release doesn't put security filters on the ASO cubes, so you have to access these cubes through Planning or through the Planning ADM driver in Hyperion Financial Reporting.  Still, ASO in Planning is a step in the right direction.  Oh, and you have to own full-use Essbase to create these ASO cubes; just a Planning license doesn't cut it.
• You can create members on the fly (kinda) if they don't exist when business rules need them.  This may only work in modules and not custom plan types.  The documentation says only modules, but I admittedly haven't tried it on custom plan types yet.
• Custom dimension hierarchies can now vary by plan type.
• Task lists can now include "Copy Version" and "Job Console" as tasks.
• There is a new Task List Dashboard (gives you an overview of tasks by user, due date, etc.) and a Task List Report Page (that exports to XLS and PDF).
• Grid Scroll Preferences lets you control if all rows/columns in a form are retrieved at once or only when you scroll down.  This should improve performance for large forms (and not slow down the internet when we want to watch cat videos on YouTube).
• Users can now control through user preferences if they want member names, aliases, or both.  This can be set to override the form settings.
• Outline Load Utility has been enhanced to do fun things like export metadata to a relational database or export data to a text file.  There's a new user interface for loading dimensions and data from flat files (and exporting them too).  This was previously only doable through the command line.
• There's a new "Plan Type Editor" that lets admins add and delete plan types to already existing applications. Previously, this required going back to the initial creation step or hacking the underlying tables.  The Plan Type Editor can be used to add ASO to an existing Planning application.
• Admins can now do some administration of Planning from within Smart View like editing dimensions, creating cubes, and refreshing cubes.
• Workforce and CapEx have been enhanced to get them up to speed with Project Financial Planning's improvements to these modules in 11.1.2.2.  You won't get these improvements with a straight migration: you have to create a new "shell" application and manually migrate your old dimensions into the new application.
• You can create, assign, and delete substitution variables directly from the Planning web interface.

Data Relationship Management

• There's an entirely new module within DRM (no word yet on if it costs money, but my gut feeling is that it's included with DRM) called "Data Relationship Governance."  To oversimplify what is actually quite cool, it adds workflow to DRM.  It lets data stewards coordinate entry, validation, and approval of hierarchies but it does a lot more too.  Users can request hierarchy changes and they go into a worklist so nothing gets lost.  It has built-in alerts for when users have requested changes and it also sends e-mails when something has been sent to you or it's something you should be informed of.  In my humble opinion, this is the greatest improvement to DRM since it was created.
• Dynamic scripting lets you use JavaScript instead of formulas to create derived properties and validations.  This is a welcome improvement over formulas since we now have access to a real programming language.
• You can now connect directly to external relational database tables to import hierarchies.

Hyperion Financial Management (HFM)

• There's a new module (that technically was released in February) called Hyperion Tax Provision.  It handles tax automation, data collection, tax calculations, reporting and it does it all within HFM.  Now when you create an application in HFM, you tell it if you want a Consolidation (traditional) application type or Tax Provision.
• Data forms let you show/hide the POV members, access the dynamic POV member lists, and run "on-demand rules" to essentially calculate the data form (by running the on-demand rule on a subset of data).
• Data grids also let you control which POV dimensions you're showing and gives you access to dynamic POV member lists.
• Admins can turn off modules for all users if they aren't applicable (for instance, if you don't allow intercompanies in HFM).

Financial Data Quality Management (FDM)

• FDM and ERPi (ERP Integrator) combine in 11.1.2.3 into a single product: FDMEE (Financial Data Quality Management Enterprise Edition).  It's more than just ERPi renamed which is what some are claiming.  Below are some of the improvements.
• The UI (user interface) is now consistent with the other Hyperion products like Planning and HFM.
• FDM is fully supported in Shared Services and Lifecycle Management.
• Data loads are sped up.  Scaling and load balancing are fully supported.
• ERPi users will now see full FDM functionality that they couldn't access before.
• SAP BW is now supported directly as a data source.
• FDM and ERPi owners have access (at no charge) to all the new FDMEE capabilities.

Financial Reporting

• I know what you're thinking and yes, Financial Reporting actually got improved in 11.1.2.3 (it is a strategic product for Oracle). So for the coolest thing you'll see since ASO in Hyperion Planning...
• Financial Reporting is now mobile.  That's right: it runs on Apple iPhones, iPads, Android phones and Android tablets.  Users can browse the repository, launch reports (HTML or PDF), run books (HTML or PDF), change POVs, change page members, expand rows and columns, and even launch related content.  Pick your jaw off the floor.
• Books have been enhanced to allow cover pages before the table of contents, embedding (and launching) Word DOCs from an HTML book, and changing the starting page number of books.
• The designer has new authoring features including a row property to show dots after the member name (to fill the column) in PDF output, custom text colors, and auto-sizing text boxes in PDF output.
• Annotations have a new auditing capability that puts annotation info (like creation date or modified date) in a log file on the server.

EPM Workspace

• OBIEE 11.1.1.7 is back in Workspace (it used to be there back in OBIEE 10).  You can both create and launch OBIEE from Workspace.
• OBIEE uses single-sign on if you're logged into Workspace (OBIEE now can also share a security model with Essbase).

Web Analysis, Interactive Reporting, SQR Production Reporting...

• Seriously?  People, move to OBIEE.  These are dead products.

General Comments

• You can upgrade to 11.1.2.3 directly only from 11.1.1.4 and 11.1.2.x.  Earlier versions require going to either 11.1.1.4 or 11.1.2.2 first.
• 11.1.2.3 still doesn't officially support Chrome or Safari.  The documentation says that Oracle recommends IE (Internet Explorer) 9 or Firefox 10 because older versions are slower.
• The EPM documentation is now fully supported on Apple mobile devices (in ePub format).  It already worked on Amazon Kindles (in Mobi format).

That's all folks.  You can thank me at Kscope13 for making it so you don't have to scour all the readme files yourself.
Oh, that reminds me. Since you read this entire essay of bullets, I owe you a reward.  When you register for Kscope13 (the best Oracle BI, EPM, and Hyperion conference in the world bar none), mention promo code IRC.  It'll save you $100 off whatever the prevailing rate is.  You can tell your friends or make them read this blog to find that out themselves.

Oracle Enterprise Performance Management 11.1.2.3 became available over the weekend, so let the rush of upgrading begin!

Here at Applied OLAP, we have already started our work with Essbase 11.1.2.3.  We have added the new version to our build list and have started testing Dodeca with Essbase 11.1.2.3.  We will make the 11.1.2.3 version of our server available to our customers within a few days.

One interesting fix that we have seen so far is that EPM 11.1.2.3 uses an upgraded version of Java, Java 1.6 update 35.   The previous EPM version, 11.1.2.2, shipped with Java 1.6 update 29 which contained a bug related to the SQL Server JDBC Driver.  The issue, known as the BEAST vulnerability, caused SQL Server JDBC connections to hang unless you made a fix to the Windows registry.  I, for one, am glad that issue is resolved.

With Oracle BI Enterprise Edition, the BI Apps and Hyperion EPM Suite, Oracle have a set of reporting and analysis tools that can be considered “best of breed” and compete with any other tool on the market. Coupled with the Oracle database, the engineered systems strategy and the Fusion Apps/Apps Unlimited ERP suites, as a BI developer it’s pretty clear we “bet on the right horse” in terms of which company we backed as developers. But it’s not always been as plain sailing as this, and like every other software company Oracle have released a few products over the years that didn’t take-off so well, never really came together or just got eclipsed by other acquisitions Oracle made over time. So lets take a light-hearted count-down through the Oracle BI, DW and ETL products released over the years that you may not have heard of, but at least some of us bear the scars from trying to implement ….

10. Oracle Warehouse Builder 10gR2′s Data Quality and Enterprise ETL Options

Back in the days before Oracle Data Integrator, Oracle Warehouse Builder was Oracle’s strategic data integration tool, initially sold on a per-developer basis but in time, bundled in with the Oracle database to match similar product packaging from Microsoft.

LIke the old saying about technology reaching the point of perfection before then becoming obsolete, in retrospect OWB10gR2 pretty much got it right in terms of Oracle-centric data integration, and the Data Quality option in particular has yet (to my mind) to be surpassed in term of data quality integration with an Oracle ETL tool. Of course, with the acquisition of Sunopsis Oracle went with ODI as their strategic ETL tool, driven mostly by the fact that it was cross-platform and had a bit more of a SOA/middleware angle than did OWB, but it’s still a shame to see OWB being de-emphasised over the years and we still miss its dimensional data modelling capabilities, integration with Oracle OLAP, and of course the data quality features that were introduced with OWB10gR2′s Data Quality Option.

9. Oracle Browser, Oracle Data Query, and Oracle Discoverer

Again, products made obsolete by newer and better ones coming through, rather than failing in themselves, Oracle Browser was Oracle’s first proper ad-hoc query tool, which in turn begat Oracle Data Query, which then begat Oracle Discoverer, still in use across many EBS sites and still with some features yet to be incorporated into Oracle BI Enterprise Edition.

But its easy to get rose-tinted-spectacles about Discoverer; having to dig out and maintain ancient Oracle JInitiator Java browser plug-ins to get Discoverer Plus to work; Discoverer Viewer only having a fraction of the functionality of Discoverer Plus; the web-based version of Discoverer first appearing with Oracle 9iAS, possibly the worst piece of Oracle software ever released, and so on. But for getting the job done with minimal fuss, Discoverer is still hard-to-beat as an ad-hoc query tool for the Oracle database, and of course its’ still available and runs now as part of the Fusion Middleware 11g setup, and it’ll still be maintained and developer for as long as there’s EBS customers out there wanting to do some simple exploration of their ERP data.

8. Oracle Data Mart Suite

Here’s one for an Oracle BI Trivial Pursuit quiz – what was the name of Oracle’s first combined ETL, data warehousing and reporting product suite, based around Oracle 8i and using ETL code licensed from Sagent? The answer is Oracle Data Mart Suite, a largely-forgotten precursor to Oracle Warehouse Builder that combined Oracle Discoverer, Oracle Designer and a tool called Oracle Data Mart Builder along with Oracle 8i to create Oracle’s first end-to-end BI & data warehousing tool.

Some of the concepts are very familiar to us now – a central repository, agents to collect metadata and run code, graphical tools to create data models and define ETL processes – but Data Mart Builder was Windows-only when Oracle were just about to move wholesale into Linux with Oracle 9i, and of course the ETL functionality pre-dates the inbuilt SQL ETL that came with Oracle 9i. Oracle Warehouse Builder 2.1 came along towards the start of the 2000s and replaced Data Mart Builder and Data Mart Designer, but Discoverer lived on and Oracle still really haven’t got a single install, single metadata store solution to replace it. Data Mart Suite even pre-dates my involvement with Oracle, but I’d be interested if anyone reading this (my guess – Kent Graziano ;-)) has any first-hand experience in working with it.

7. Oracle Darwin

Just around the same time as Oracle 9i was released, Oracle made the acquisition of Thinking Machines, a data-mining and supercomputer company based out in the States who sold a product called Darwin, a Clementine-like GUI analytical workbench that Oracle later rebranded as “Oracle Data Mining”.

Darwin never really saw the light of day with Oracle but the internal algorithms and technologies went on to form the core of the Data Mining Option for the Oracle Database (now part of the wider database Advanced Analytics Option), which now has GUI elements of its own but does all of the processing in the Oracle database. Technology derived from Darwin can also be found today in products like Oracle Spend Classification, working under the covers to classify corporate spend using data mining classification algorithms.

6. Oracle BI Applications 7.9.5.2

Now we’re getting into the interesting ones. Four weeks of my life I’ll never get back were spent back in 2009 getting to grips with the first version of the Oracle BI Apps that used ODI, rather than Informatica, as the ETL tool. Sources and targets in this initial release were limited to just EBS 11.5.10 on Oracle Database 10gR2, but other than that it was fully-functional, with a method for doing customisations, new configuration tools that did away with the need for the DAC, and all of the relevant SDE and SIL mappings re-implemented as ODI interfaces and packages.

But this was back in the days of ODI10g, and there were no load plans or any other features since introduced to ease the move to ODI with the BI Apps, and the customisation approach was slightly scary and complex, to say the least. In the end, only one release of BI Apps on ODI10g ever came out, but of course we’re now just nearing the point where BI Apps 11g gets full support for ODI as an alternative to Informatica as the ETL tool, and seeing the various workarounds Oracle had to do with ODI to get it to work as an Informatica alternative back in 2009 made it obvious to me why features such as load plans were introduced over the past few years.

5. The Oracle BI Spreadsheet  Add-in for Excel

A spreadsheet add-in with less useful functionality than Oracle BI Office, the GUI performance of Warehouse Builder and an OLAP server back-end that nobody used. One of a number of entries in this chart based around Oracle 9i OLAP.

This, and No.1 on our list were probably the single biggest reason Oracle ended-up buying Hyperion – the chances of an accountant actually using this Excel add-in, as opposed to say Hyperion’s Essbase add-in, were about as close to zero as you could get, assuming you could find anyone still using Oracle OLAP after the bodged migration from Express Server. But – Oracle 9i OLAP going so badly paved the way, in time and several years later, for the Hyperion acquisition, and now Oracle OLAP sensibly focuses on the Simba MDX Provider for Oracle OLAP along with Microsoft Excel pivot tables, the descendants from this product release are actually pretty darn good.

4. Oracle Business Intelligence Essbase Integrator

Another one for Oracle BI Trivial Pursuit – which product from Oracle integrates OBIEE, Essbase and the BI Apps, so that you can click on links on your BI Apps dashboard and launch Smarview, preserving the POV context from the related analysis on the dashboard?

The answer is the Oracle Business Intelligence Essbase Integrator, a product launched by Oracle back in 2010 and which appeared to me, at the time, as wildly ambitious but solved a real problem – how do you combine the capabilities of Essbase and OBIEE whilst in the background, keeping their metadata in-sync.

Rather predictably, we didn’t really hear much about this product again which was a shame, as the concept was superb (albeit a bit “Heath Robinson”, or “Rube Goldberg” as the Americans would say). I suspect we’ll see something around this same concept going into the 12c timeline as it’s such an obvious “win” for Oracle, but for now, it’s a product best known as the challenge we set Stewart Bryson back at the Brighton BI Forum 2011 when Venkat couldn’t make it over – present Venkat’s OBI Essbase Integrator session as if it were his, and without seeing the content of each slide until it came up on the projector.

3. Oracle Daily Business Intelligence

Prior to the BI Apps and around the time of Oracle E-Business Suite 11i, reporting against EBS data was done either through Discoverer, or through Oracle Reports and FSGs. Oracle Discoverer could integrate with Oracle Portal, but it was fairly primitive integration and Portal wasn’t really cut-out to be a BI Portal, more being used for intranet-style applications like data entry and staff directories.

Oracle Daily Business Intelligence (DBI) improved on this situation in two ways; first, it was based off-off Oracle Database materialised views, speeding up report response times and, in theory, taking the load off of your underlying ERP system; second, it had its own custom portal and dashboard framework that was faster, more responsive and “tighter” than Oracle Portal and its Discoverer portlets.

DBI may well have got more traction over time but probably was eclipsed by the major acquisition spree that Oracle went on in the early 2000′s, buying Peoplesoft (and thereby JD Edwards) and Siebel, and then using Siebel Analytics along with BEA’s WebLogic tooling to create the successor BI-on-ERP platform, OBIEE and the BI Apps. Which was probably a good thing in the end, as I never met anybody actually able to customise Daily Business Intelligence, a task that makes customising the BI Apps seem like plugging your iPod into iTunes running on a Mac.

2. Oracle 9i OLAP, and Oracle Business Intelligence Beans

On to the final two, and they’re both based around Oracle 9i OLAP, Oracle’s (in retrospect) disasterous replacement for Oracle Express Server which lost them the OLAP market they owned back in the late 90′s. I’ve got mixed feelings on 9i OLAP as I know many of the people involved in its development and marketing, and nowadays in its database OLAP option form it’s an excellent product, I think technically and architecturally better than Essbase. But the introduction of Oracle 9i OLAP was a masterclass in suicidal product marketing; first, there was no backward compatibility with Express Server tools, so all of the users of Oracle Financial Analyzer and Oracle Sales Analyzer had to wait years for the product at #1 in our list, with of course most of them decamping to Microsoft OLAP Services or Arbor Essbase instead. The first version of Oracle 9i OLAP was ROLAP only, with terrible, buggy performance and a wait of a year or so before the MOLAP version came out, again without backwards compatibility with Express Server tools (due to the removal of the key SNAPI and XCA communications APIs that Express Server applications used)

 

All of this made great technical sense, and if the MOLAP version of 9i OLAP had come out at the same time as the ROLAP version, and if Oracle somehow managed to keep SNAPI support so that OFA and OSA customers could technically migrate their OLAP cube to Oracle 9i without loosing their tool access. It may have all worked out. But the “icing on the cake” was the lack of any proper ad-hoc or OLAP query tool support right at the start (Discoverer “Drake” came a few years later), with customers expected to – get this -write their own BI tool using a Java component technology called Oracle Business Intelligence Beans.

A few parters, including myself in this bit of history from around ten years ago, gamely tried to generate interest around 9i OLAP and BI Beans, but combined with Microsoft’s entry into the OLAP market and Arbor (and then Hyperion’s) focus on the finance department, rather than DBAs and IT who never actually buy OLAP servers, Oracle OLAP never regained the market share that Express Server had, even though as I said earlier it’s arguably a better, more scalable and easier-to-manage OLAP Server than Essbase.

The last laugh is on the BI Beans product development team though, as the BI Beans query builder became the inspiration for OBIEE 11g’s “Selection Steps” feature, whilst its data visualisation components found their spiritual successor in ADF’s Data Visualization Tools (DVT) feature, which provides the BI visuals behind OBIEE, the latest version of Endeca Information Discovery, and of course the Oracle Fusion Apps.

1. Oracle Enterprise Planning & Budgeting

Number one in our list of Oracle’s slightly crazy BI tools from the past was Enterprise Planning & Budgeting, the long-awaited replacement for Oracle Financial Analyzer and Oracle Sales Analyzer based around the Oracle 9i OLAP platform. More akin to the Stone Roses’ “Second Coming” and about as well critically received, EPB was the “aircraft carrier” to OFA’s “motor torpedo boat”, had a list as long as your arm of critical patches you had to apply before you could use it, and required installation along with EBS (and knowledge of a set of arcane setup steps) before you could use it.

Coupled with a painfully-slow user interface for users typically used to split-second OFA response-times, EPB was long in the coming but quickly despatched when Oracle first adopted Siebel Analytics and the BI Apps as their new BI Platform, and then bought Hyperion and made Essbase and Hyperion Planning the centrepiece of their performance management strategy, something that carries on to this day.

So there we have it – a light-hearted look through some of the Oracle BI products that didn’t make it to the big time, and a bit of history to explain why OBIEE and EPM Suite are the tools we use today. Most of the Oracle PMs who looked after these tools are still with us, working on OBIEE and its related technologies, so apologies if I’ve inadvertently offended anyone by picking on one of the products they looked after – it was all fun at the time and most of the products would have stayed with us, and gone on to be successes were it not for the massive strategic shift Oracle made back at the turn of the century towards cross-platform, and away from the Oracle database begin the centre of everything. Let me know if you’ve had any experiences with these tools, or if you’ve got any corrections or additions to their stories.

In the previous three articles in this series (and well done for everyone that’s still with us), we looked at how OBIEE 11.1.1.7 and ODI 11.1.1.6+ can access Hadoop data sources via a related technology called Hive, looking in the second article in more detail at the OBIEE aspect including how the OBIEE metadata import was set up, and how new HiveODBC drivers supplied by Oracle make this all possible. In the last posting in the series, we saw how ODI can be used to populate the Hive “warehouse” tables, stored on Hadoop’s HDFS (Hadoop Distributed File System) storage grid, that are then analysed via HiveQL and MapReduce, and in this final post, we’ll take a look at how ODI can go beyond simple loading from file into Hive tables and start to do some of the data transformation, and integration tasks that we regularly perform with more traditional data sources and targets.

In the previous ODI and Hive/Hadoop example, we used an ODI interface to load data from a local file into a Hive table, using the IKM File to Hive knowledge module. In this first example columns in the source file were mapped 1:1 into “columns” in the Hive table source, like this:

whilst in the second example, we used a Hadoop “SerDe” serializer-deserializer transformation to parse incoming weblog rows into the target Hive column format, again using this same IKM File to Hive knowledge module.

Which of course is great if you’re loading data directly from files into corresponding Hive warehouse tables, but what if you’ve already got data in Hive but you want to “reshape” or transform it, creating the equivalent of a star schema, say, of a set of “normalised” tables?

Hive, compared to Oracle at least, is pretty simplistic in terms of the data manipulation you can do with it, and is more akin to working with Oracle external tables than a full insert-update-delete-ACID-compliant database (but then again, it was never intended to be that). You can only insert new data into Hive tables, not (ordinarily) update rows or delete them, with the only way to delete data in a Hive table being to drop it, then re-create it new. HiveQL has syntax for joins, group by and some functions, and you can drop in your own custom MapReduce scripts for the more complex stuff – which is not bad going and probably more than enough for your average Hadoop/Big Data analyst.

But the scenarios we’re considering probably won’t have a Hadoop expert around, and are typically something like:

• We’ve got some marketing or analyst-type users who want to access some behavioural, weblog or activity data sitting in Hadoop/HDFS/Hive, they’re happy users of OBIEE, and we want to be able to connect OBIEE to this data so that they can report on it – which probably means “reshaping” it somewhat to fit OBIEE’s preference for star-schema (or at least simplified, denormalized) source data models, or
• We’ve got some data sitting in Hadoop/HDFS that we’d like to add in as a new source into our data warehouse, and ODI is the preferred tool for bringing in new sources

In neither case do we have much in the way of Hadoop or MapReduce skills, so we can either (a) use OBIEE’s BI Repository to do some source data reshaping, or even better (b) do the job properly, maybe as part of a much wider ETL process, using ODI. The key thing is fitting Hadoop into ODI’s way of doing things, and giving ODI the ability to do Hadoop and Hive-specific tasks through its extensible “knowledge module” framework.

In the example above that I used to illustrate ODI’s ability to work with Hive, I used the IKM File to Hadoop knowledge module that comes as part of the ODI Application Adaptor for Hadoop (ODIAAH); ODIAAH is one of a number of Fusion Middleware “application adapters”, and is licensed separately (but in conjunction with) ODI EE costing around the same as GoldenGate (around $17000/processor, according to the current tech price list). For some more background into ODIAAH this self-study training session goes into the basics, but at a high-level this adapter actually ships four mainly Hive-based ODI KMs that enables data loading, transformation and integration between Hadoop/Hive/HDFS and the Oracle database (amongst other sources/targets);

• IKM File to Hive (Load Data) : what we’ve been using so far, used to load Hive tables from local and HDFS-stored files
• IKM Hive Control Append : used for loading data into Hive tables, with data sourced from one or more other Hive tables – the equivalent of loading one relational table by joining, filtering and transforming other tables
• IKM Hive Transform : a variation on the previous KM that allows incoming data to be transformed via Python or Perl scripts
• IKM File-Hive to Oracle : load data from either a Hive table, or from files stored on HDFS, into an Oracle database using the separately-licensed Oracle Loader for Hadoop
• CKM Hive : a check knowledge module for Hive (which because of its “schema on read” rather than “schema on write” approach, doesn’t natively support keys or constraints)
• RKM Hive : a “reverse-engineering” KM that allows ODI to read table definitions from the Hive metastore, and use them to create ODI datastore definitions

A typical Oracle BI&DW project is going to use these KMs for a couple of main reasons; one, as stated before, is because their might be information sitting in Hadoop that the BI/DW system wants access to, and would otherwise have to start writing MapReduce, Pig, Sqoop etc code to get into their data warehouse. The other might be to leverage Hadoop’s ability to crunch and count large sets of data massively parallel, at relatively low cost, with ODI then initiating and monitoring the process, then loading the results into a more traditional Oracle data store. Other technologies, some implemented or licensed by Oracle, also surround this process – Oracle NoSQL database for example – but for now lets concentrate on Hadoop and Hive, and see what else these KMs can do for us.

Starting off with IKM Hive Control Append, you might have a situation where you’ve already got some data in Hive tables (perhaps through loading them up using ODI’s File to Hive (Load Data) KM, but the data needs joining, transforming, filtering or otherwise reshaping before you can connect a tool such as OBIEE to it. In this case it doesn’t make sense to use ODI File to Hive (Load Data) as this KM is for when the data sits outside Hive in source files, so instead we can use IKM Hive Control Append to truncate/insert append new rows into an existing target Hive table.

This process works very similar to regular table loading IKMs (except of course you can’t insert/update into the target, only insert append); for example, in the screenshot below, two Hive tables, one for customer information and one for salesperson information, are joined and the results transformed using HiveQL with the results loaded into another Hive table.

Looking at the loading code generated by ODI for this interface, you can see that the HiveQL used to extract and load the data looks very similar to Oracle SQL, and in fact this KM is about the closest one to “regular” relational database ones out of the set of ODIAAH knowledge modules.

Things get more interesting with the IKM Hive Transform, a knowledge module that takes data from any source and loads it, via custom shell scripts, into a target Hive table. These transformation shell scripts are typically written in Python or Perl, and give you the ability to write your own custom pre-processing or transformation code that (via ODI’s temporary interfaces feature) can then be used as “inline views” or multi-step processes when performing more complex data loading processes around Hive. In the example below, a three-step data loading process first creates and then loads an Apache Weblog file using IKM File to Hive (Load Data), then sessionizes (tags the log file with IDs to identify all activity within a particular browser session) using IKM Hive Transform:

Taking a look at the final “Sessionize Weblog” step in more detail, you can see in the interface Flow tab that this step uses the IKM Hive Transform module, and a script called “sessionize.pl” to do the transform work.

The perl script itself then parses through the log file information and works out the start, and stop points for each individual user session, outputting the results which are then transformed into the correct target columns by the PRE_TRANSFORM_DISTRIBUTION settings in the KM options.

Finally, the IKM File-Hive to Oracle knowledge module takes things in the other direction, extracting from Hive tables or HDFS files into an Oracle database, via the Oracle Loaded for Hadoop big data connector, but that’s really a topic in itself and for another day, when I’d like to look in more detail at the big data connectors in general, and how you can leverage Hadoop and HFDS from within SQL and PL/SQL commands. For now though, this concludes my look at Hadoop connectivity from within OBIEE 11.1.1.7 and ODI.11.1.1.6, but if you’re like me this brings up as many questions as it answers; for example:

• How would I go about setting up my own Hadoop/Hive/HDFS development environment, and can I use Windows or does only Linux make sense?
• Just what could I do with Hadoop as a data source that I can’t do with regular Oracle and file sources; in terms of scale, and also complexity/speed?
• Can ODI also make use of Hadoop data loading/transformation tools like Pig, or Sqoop?
• Do any of the related/commercialized Hive/Hadoop technologies add anything to these scenarios – for example, Cloudera’s Impala (for BI metadata/reporting) or Cloudera Manager (for Hadoop administration)?
• And – is there a time/volume threshold where Hadoop makes more sense as a data processing platform than an Oracle database?

If these sound interesting to you, they’re exactly what I’ll be covering during my part of the Data Integration Masterclass at the Rittman Mead BI Forum 2013 events in Brighton and Atlanta, running in just a few weeks time. We’ve still got a few places left, so if you’re interested and want to see all this technology in action, sign-up now and I’ll hopefully see you soon.

In the first two parts in this series, I looked at the recently-added support for Apache Hadoop as a data source for OBIEE 11.1.1.7 and ODI 11.1.1.6, and explained how the Hadoop support was really enabled through a related technology called Hive. In the second part in the series I showed how OBIEE 11.1.1.7 could report against “big data” sources using Hadoop and this Hive technology, but this all of course pre-supposes that we have data in Hive in the first place. So what actually is Hive, how do you load data into it, and can ODI help with this process?

To take a few steps back, Apache Hive is a Hadoop-family project that provides a “data warehouse” layer over Hadoop, through a metadata layer not unlike OBIEE’s RPD together with a SQL-like language called HiveQL. Coupled with ODBC and JDBC database drivers, BI tools like OBIEE use Hive to get access to big data sources, as the HiveQL language that Hive uses is very similar to SQL used to access databases such as Oracle, SQL Server or mySQL. Delving a bit deeper in the Hive product architecture, as shown in the diagram below Hive has a number of components including a “database engine”, a metadata store, APIs for client access, and a link through to Hadoop to actually load, process and retrieve data in HDFS (Hadoop Distributed File System).

So what’s HDFS then? HFDS is a fault-tolerant, distributed filesystem that’s a core part of Apache Hadoop, and stores the data that MapReduce jobs then process via job trackers, task trackers and all the other Hadoop paraphernalia. HDFS is accessed through a URI (URL) rather than through your Linux filesystem browser, but distributions such as Cloudera’s CDH3 and CDH4 ship with tools such as Hue, shown below, that provide a web-based interface into HDFS so that you can browse HDFS like a regular OS-level filesystem.

Notice how there’s a “user” folder like we’d get with Linux, and within that folder there’s a home folder for Hive? With Hive, generally the data you manage using Hive is actually loaded into a directory structure under the “hive” user, either using data taken from another directory area in HDFS or from external files. Hive’s data is still in file form and accessed via MapReduce and Hadoop, but it’s in a directory area away from everything else. You can, however, tell Hive to create tables using data held elsewhere in HDFS, analogous to Oracle’s external tables feature, which then skips the data loading process and just maps table structures onto files held elsewhere in the Hadoop filesystem.

In most cases when we’re considering OBIEE accessing Hadoop data via Hive, the data would have been loaded into Hive-mananged tables tables beforehand, though it’s possible that Hive table metadata could have been mapped onto other data in HDFS. In your own particular Hive implementation and assuming you’ve got Hue installed, and Beeswax, a table browser for Hive that usually comes with Hue, you can see where each individual table within your Hive metastore is actually held; in the examples below, the dwh_customer Hive table is a managed table and has its data stored within the /user/hive/warehouse/ HDFS directory, whilst the ratings table has its data stored outside of Hive’s directory structure, but still within the HDFS managed filesystem.

So how does one create a Hive table, load data into it and get it ready for OBIEE access, and can ODI help with this, as we asked earlier? Before we get into ODI then, let’s take a look at how a Hive table is created and loaded, and then we’ll see how ODI does the same job.

With thanks to the ODI product development team’s David Allan, who put together some great Hive and ODI examples in this blog post, let’s start by creating a Hive table against the same movie ratings data in the right-hand screenshot below, but this time with the data actually loaded into Hive’s directory structure (i.e. a “managed” table). From the Hive command-shell, I type in the following commands to create the managed table, after SSH’ing into the VM running Hive:

officeimac:~ markrittman$ ssh oracle@bigdatalite
Warning: Permanently added the RSA host key for IP address '192.168.2.35' to the list of known hosts.
oracle@bigdatalite's password:
Last login: Mon Apr 22 10:59:07 2013 from 192.168.2.47
=====================================================
=====================================================
Welcome to BigDataLite
run startx at the command line for X-Windows console
=====================================================
=====================================================

Host: bigdatalite.us.oracle.com [192.168.2.35]

[oracle@bigdatalite ~]$ hive
Hive history file=/tmp/oracle/hive_job_log_oracle_201304250732_1523047910.txt

hive> create table movie_ratings (user_id string
> , movie_id string
> , rating float
> , tmstmp string)
> row format delimited fields terminated by '\t';

OK
Time taken: 3.809 seconds
hive>

At this point the table is created but there’s no data in it; that part comes in a moment. I can see the table structure and its empty state from the Hive command-line:


hive> describe movie_ratings;
OK
user_id string
movie_id string
rating float
tmstmp string
Time taken: 0.168 seconds

hive> select count(*) from movie_ratings;
Total MapReduce jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=
In order to set a constant number of reducers:
set mapred.reduce.tasks=
Starting Job = job_201303171815_0021, Tracking URL = http://localhost.localdomain:50030/jobdetails.jsp?jobid=job_201303171815_0021
Kill Command = /usr/lib/hadoop-0.20/bin/hadoop job -Dmapred.job.tracker=localhost.localdomain:8021 -kill job_201303171815_0021
2013-04-25 07:40:51,581 Stage-1 map = 0%, reduce = 0%
2013-04-25 07:40:56,617 Stage-1 map = 0%, reduce = 100%
2013-04-25 07:40:58,640 Stage-1 map = 100%, reduce = 100%
Ended Job = job_201303171815_0021
OK
0
Time taken: 12.931 seconds
hive>


and also from the Beeswax web UI:

So how do we get the data into this table, without any tools such as ODI? I can either load data straight from files on my local workstation, or I can upload them, for example using Hue, into the HDFS filesystem first.

Now I can use the HiveQL LOAD DATA command to load from one of these HDFS tables into Hive, and then count how many rows have been loaded, like this:


hive> load data inpath '/user/oracle/movielens_src/u.data'
> overwrite into table movie_ratings;
Loading data to table default.movie_ratings
Deleted hdfs://localhost.localdomain/user/hive/warehouse/movie_ratings
OK
Time taken: 0.341 seconds

hive> select count(*) from movie_ratings;
Total MapReduce jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=
In order to set a constant number of reducers:
set mapred.reduce.tasks=
Starting Job = job_201303171815_0022, Tracking URL = http://localhost.localdomain:50030/jobdetails.jsp?jobid=job_201303171815_0022
Kill Command = /usr/lib/hadoop-0.20/bin/hadoop job -Dmapred.job.tracker=localhost.localdomain:8021 -kill job_201303171815_0022
2013-04-25 08:14:24,159 Stage-1 map = 0%, reduce = 0%
2013-04-25 08:14:32,340 Stage-1 map = 100%, reduce = 0%
2013-04-25 08:14:42,420 Stage-1 map = 100%, reduce = 33%
2013-04-25 08:14:43,428 Stage-1 map = 100%, reduce = 100%
Ended Job = job_201303171815_0022
OK
100000
Time taken: 26.32 seconds
hive>


So how does this process look when using ODI to do the Hive data loading? Let’s start with importing the Hive table metadata for the movie_ratings table I just created from the Hive command-line shell, by going over to the Topology navigator in ODI 11.1.1.6 – note that you’ll need to configure ODI to connect to your Hive, HDFS and Hadoop environment beforehand, using the Oracle Data Integrator for Hadoop documentation as a guide, with this adapter being an extra-cost license option on top of base ODI Enterprise Edition.

Hive has its own technology type within the Topology navigator, and you create the connection through to Hive using the HiveJDBC driver, first adding the connection to the Hive server and then specifying the particular Hive database / namespace, in this case selecting the “default” database for my Hive system.

Now I can reverse-engineer the Hive table structures into a Designer navigator model, just like any other relational table structure.

Within the ODI Topology navigator you can then create File technology connections either to files held in HFDS, or more likely with ODI to files on your workstation, or server, filesystem, like this:

and then add the filedata stores to the Designer Navigator Model list, entering the correct delimiter information and reversing the column definitions into the datastore definition.

Now it’s a case of creating an interface to load the Hive table. In this instance, I map each of the source file “columns” into the Hive table’s columns, as the source file is delimited with an easily-usable structure.

Then, over in the Flows tab for the interface, I make sure the IKM File to Hive knowledge module is selected, keep the default values for the KM options (more on these in a moment), and then save the interface.

Now it’s a case of running the interface, and checking the results. Notice in the Operator navigator code panel, the LOAD DATA command that ODI is generating dynamically, similar to the one I wrote manually earlier on in the article.

Going back to my Hive command-line session, I can see that there’s now 100,000 rows in the movie_ratings Hive table.


hive> select count(*) from movie_ratings;
Total MapReduce jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=
In order to set a constant number of reducers:
set mapred.reduce.tasks=
Starting Job = job_201303171815_0024, Tracking URL = http://localhost.localdomain:50030/jobdetails.jsp?jobid=job_201303171815_0024
Kill Command = /usr/lib/hadoop-0.20/bin/hadoop job -Dmapred.job.tracker=localhost.localdomain:8021 -kill job_201303171815_0024
2013-04-25 16:59:12,275 Stage-1 map = 0%, reduce = 0%
2013-04-25 16:59:18,346 Stage-1 map = 100%, reduce = 0%
2013-04-25 16:59:29,467 Stage-1 map = 100%, reduce = 33%
2013-04-25 16:59:30,475 Stage-1 map = 100%, reduce = 100%
Ended Job = job_201303171815_0024
OK
100000
Time taken: 27.251 seconds


Now in many cases the data going into a Hive table isn’t neatly arranged into columns within delimited files; it could be, for example, web log data that you’ll need to parse using regular expressions or other APIs or standard parsers. When that’s the case, you can use an option with the IKM File to Hive knowledge module to override the normal column-to-column mappings and instead use an expression, something Oracle have done in their demo environment for parsing these types of log files.

“ROW FORMAT SERDE” is a reference to Hive’s “Serializer – Deserializer”, or row-formatting feature, that gives you the ability to use regular expressions and other data manipulation techniques to, in this case, allocate incoming file data to the proper columns in the target hive table.

So now we’re at the point where we can use ODI to populate the Hive tables that OBIEE in turn uses to access Hadoop data sources. But what if the data we want to load into Hive isn’t in the format or shape we need, and we need to join, filter or otherwise work with Hive data and tables before we can report on it. And what if we want to get data out of Hive and into regular tables if a relational data store makes more sense than Hadoop, for a particular reporting requirement? Check back tomorrow for the final part in this series, where we’ll answer these remaining questions.

Apart from system-wide improvements to Essbase integration across OBIEE 11.1.1.7, the other Essbase-related improvement that came with this latest release was the (re-)introduction of SmartView as the replacement for Oracle BI Add-in for Microsoft Office (“BI Office”), OBIEE’s previous MS Office solution. As a reminder, BI Office appeared with OBIEE 10.1.3.3 back in 2007/8 and supported integration with Microsoft Excel and Powerpoint, allowing you to download analysis views from the BI Catalog and then view them within Excel and Powerpoint. 

What you couldn’t do with BI Office though was use it to create new analyses, or upload what you’d created back to the BI Catalog. There was also no integration with Microsoft Word or Outlook, which meant it was a feature meant more for viewing and copying analyses into Excel and Powerpoint rather than as a “first class” report authoring environment.

Then when OBIEE 10.1.3.4 was released, a number of EPM Suite products were integrated with OBIEE, including Workspace (now resurrected with OBIEE 11.1.1.7), SmartSpace (where did that go?) and SmartView, the long-term replacement for Essbase’s somewhat minimalist Excel Add-in. This was all good stuff except that, in terms of OBIEE support, this version of SmartView was essentially unusable, rendering OBIEE data in an Essbase-like way that made little sense for an OBIEE user.

“The UI takes a bit of getting used to” was my comment at the time, which in retrospect was a bit of an understatement and this version of SmartView had little to no take-up within the OBIEE world, with BI Office carrying on until now as the only viable MS Office integration approach. Now though, the new improved version of SmartView is with us, so how well does it work with OBIEE data?

SmartView can be download from the BI Presentation Services homepage, but note that this is the 32-bit version and you’ll need to go to My Oracle Support for the 64-bit version, available using patch ID 16238382 (at the time of writing, for SmartView version 11.1.2.2.310). Once its installed, select SmartView > Options > Advanced and enter your general EPM Suite Smartview Provider Services URL into the Shared Connections URL setting (in the format http://[machine_name:port}/workspace/SmartViewProviders), like this:

This setting only covers SmartView connecting to Essbase and Financial Reporting, so to connect to OBIEE's Presentation Services Catalog you'll need to create what's called a Private Connection (or define a shared connection for OBIEE within an XML file, as detailed in the SmartView 11.1.2.2.310 docs), by pressing the Panel button in the menu ribbon, selecting Private Connections from the Smart View menu, then clicking on the Create new connection button.

Then, when prompted for the SmartView connection type, select Oracle BI EE, then type in the OBIEE SmartView URL in the format http://[machine_name:port]/analytics/jbips, and press Finish to complete this part of the process.

Then, when prompted enter the username and password for your OBIEE system, and then save the connection as a private connection to your workstation.

Now you should be able to browse the BI Catalog and select a SmartView report, for example, to view within Excel.

Or you can select any view from a regular analysis, and add that to Excel just as you did with BI Office.

More importantly though, the View Designer feature allows you to create a new report from scratch, selecting from any subject area in the BI Catalog and creating a report from right within Excel.

This report can then be manipulated either as an Excel pivot table (pictured below) or an OBIEE pivot table, giving you an OBIEE-within-Excel experience far more intuitive and usable than the earlier incarnation of SmartView.

Additional calculated fields can be added, in what is arguably a more obvious way than you’d do so in the Analysis Editor…

… and charts can be developed as well, using a similar set of of chart types to the ones provided by the Analysis Editor.

Then, once you’re done, you can either save the Excel (or Word, or Powerpoint, or whatever) document to your workstation’s filesystem, or you can upload to the BI Presentation Catalog using the Publish View button…

… and then – get this – open the report in the Analysis Editor, just like any other analysis in the catalog. Impressive stuff (although the calculation defined in Excel didn’t make it through to OBIEE, and the upload feature only seems to bring a single view at a time, but this is version 1.0)

There’s tons more to Smartview and in reality, presumably some of the new OBIEE stuff won’t work properly in this first release, but it’s a huge improvement over the old OBIEE MS Office plug-in, and it’s also useful being able to use the same MS Office plugin for all Oracle’s BI & EPM tools, with full 32 and 64-bit support for all the modern MS Office versions.

One of the major new feature areas in OBIEE 11.1.1.7, but which has so far got very little attention, is the significant improvement in integration between Essbase, the Hyperion EPM Suite, and OBIEE 11g. The integration between EPM Workspace and OBIEE’s Presentation Services which disappeared when 11g came along is now back, along with installation and security integration, a new version of SmartView that (properly) supports OBIEE as a data source, and the ability to spin-off aggregates from the RPD into Essbase ASO cubes.

Now some of these features of course made an appearance in the earlier, 11.1.1.6.2 BP1 release, and integration between OBIEE 11g and EPM Suite has been happening on-and-off right back from the OBIEE 10g days, but where we’re at now with OBIEE 11.1.1.7 is the delivery of a number of things that customers have long been asking for, including:

• The ability to run OBIEE from within EPM Workspace, with single sign-on between the two
• Shared security provisioning and organisation between Essbase and OBIEE, through application roles and policies
• The ability to install Essbase and the other EPM tools into the same WebLogic domain as OBIEE, using a single installer
• A proper Excel (and Word, Powerpoint, Outlook) add-in for OBIEE, with the ability to author reports as well as run existing Answers-authored ones

This is actually one of a number of new feature areas that came with 11.1.1.7 that have had little publicity; as well as better Essbase integration, there’s actually now support for multi-tenancy in the RPD and catalog, Hadoop integration (which we covered in a blog post last week), the View Suggestion Engine, the inevitable changes to MUD, and quite a few others, some of which I’ll try and cover in the next few days and weeks, but for now let’s look at these new Essbase/EPM integration improvements, starting with installation of Essbase and its related tools into the OBIEE WebLogic domain.

As I mentioned back in my OBIEE 11.1.1.7 New Features posting a few weeks ago, the OBIEE product installer now offers Essbase as an installation option alongside OBIEE, Real-Time Decisions (RTD) and BI Publisher. As with RTD, Essbase isn’t included in the base OBIEE+ license, but it is included in Oracle BI Foundation Suite, the product package that Oracle encourage new customers to take out an includes OBIEE, Scorecard & Strategy Management, Essbase and BI Mobile. Selecting Essbase during the install process installs it, and the other EPM Suite tools, in the same WebLogic domain as OBIEE, and you can see Essbase within Fusion Middleware Control as a product – separate from OBIEE – that you can manage and monitor.

Essbase Server, and Essbase Studio (the client/server tool used to design and build Essbase cubes) are also now controlled and monitored through OPMN, something that’s been a feature of EPM Suite for several releases now but which is, of course, new for OBIEE.


[oracle@obiee11117 ~]$ cd /home/oracle/obiee/instances/instance1/bin
[oracle@obiee11117 bin]$ ./opmnctl status

Processes in Instance: instance1
---------------------------------+--------------------+---------+---------
ias-component | process-type | pid | status
---------------------------------+--------------------+---------+---------
essbasestudio1 | EssbaseStudio | 12682 | Alive
essbaseserver1 | Essbase | 12685 | Alive
coreapplication_obiccs1 | OracleBIClusterCo~ | 12686 | Alive
coreapplication_obisch1 | OracleBIScheduler~ | 12687 | Alive
coreapplication_obijh1 | OracleBIJavaHostC~ | 12683 | Alive
coreapplication_obips1 | OracleBIPresentat~ | 12684 | Alive
coreapplication_obis1 | OracleBIServerCom~ | 12689 | Alive

[oracle@obiee11117 bin]$


So something that’s been an issue for EPM customers upgrading from OBIEE 10g to 11g was the removal, at the time, of the ability to integrate OBIEE’s Presentation Services within EPM Workspace, and the SSO link between the two products. Back with OBIEE 10.1.3.4 there was an admittedly complicated but supported and working process to integrate the two products together, allowing EPM Workspace customers to “skin” OBIEE to look like Workspace and run the two products together, albeit with separate report catalogs, security models and so forth.

This, coupled with the removal of OBIEE’s Hyperion custom authenticator for the RPD left many EPM Suite customers upgrading to OBIEE 11g in the lurch, leading to workarounds such as this one that we put together recently for one of our customers. Well this integration (mostly…) is back with OBIEE 11.1.1.7, so let’s see what it does, and what functionality is still missing compared to OBIEE 10g.

First off, Essbase and EPM Suite as installed as part of an OBIEE installation isn’t quite the same as EPM Suite installed standalone; most importantly, Essbase in this OBIEE incarnation has a different security model than “standalone” EPM Suite, in that it uses the same system of application roles and policies that the Fusion Middleware 11g-centric OBIEE 11g does, rather than the Shared Services and groups that standalone EPM Suite does. Also, the OBIEE 11.1.1.7 install installs just the following EPM Suite products:

• Essbase Server, including Essbase Agent, Essbase Studio, Essbase Administration Services, Provider Services
• Financial Reporting
• Calculation Manager

Therefore you don’t get Planning, Web Analysis and so forth, and you can’t subsequently install them into the domain and Fusion Middleware Control afterwards – so think of Essbase and the EPM Suite tools in this context as an add-on and complement to OBIEE, not a full installation of EPM Suite in their own right. Moreover, the majority of Essbase administration tasks which for standalone EPM Suite installs are performed through MaxL, Shared Services and EAS are performed through Fusion Middleware Control, and Essbase high-availability and clustering works different within this context, for example. The standard product architecture diagram for OBIEE and Essbase combined within the 11.1.1.7 release therefore gets updated, with a number of products added to the Java components, and System components part of the diagram, like this:

Now, when installed as part of OBIEE 11.1.1.7′s WebLogic domain, EPM Workspace is available at http://[machine_name:port]/workspace, and when you launch it you’re presented with a view into the BI Catalog, and menu options to administer the various EPM and BI tools from one place. 

Within this catalog are both OBIEE objects such as analyses, dashboards and agents, and EPM objects such as Financial Reporting and SmartView reports.

There are limits to this EPM/BI Catalog integration though – FR reports, for example, can only be opened using the File > Open dialog in EPM Workspace, with an error message showing if you just click on the report itself in the BI Catalog view within EPM Workspace. But SSO between Workspace and OBIEE seems to work (as in, you don’t need to re-enter your BI password when clicking on an analysis in the Workspace Catalog view) as both OBIEE and EPM are working off of the same Fusion Middleware security model, which (the lack of) explains why the feature disappeared for so long after OBIEE 11g was introduced.

Now that OBIEE and Essbase share the same security, the need for the old HSS Custom Authenticator has now gone away, though of course this will only be of use if a customer has moved their Essbase installation into the OBIEE domain, with standalone EPM Suite installations still needing the security workaround mentioned earlier in this article. There’s no upgrade path from standalone EPM Suite installations to this integrated arrangement, so most probably any users of Essbase within this new 11.1.1.7 context will be installing it “net-new”, with the main objective being to enhance their existing BI setup rather than merging their separate BI and EPM platforms into one. 

As you’ve probably picked-up by now, much of this new integration ability is down to security harmonised across both Essbase and OBIEE, or more accurately Essbase now having an option to use Fusion Middleware 11g security rather than Hyperion Shared Services. So what does Essbase and FMW11g security look like in practice? Let’s head over to Fusion Middleware Control, in particular the Application Policies administration screen, to take a look.

The big difference when Essbase runs as part of an Oracle BI domain is that authentication, and authorization for Essbase use Fusion MIddleware security rather than Shared Services or Native Essbase security. Although Essbase Administration Services ships with OBIEE 11.1.1.7, you should use Fusion Middleware Control to enable access to particular Essbase databases, and give permission to access tools such as Financial Reporting or Administration Services; the only security role for EAS and MaxL in this setup is to create the Essbase data and metadata filters; these filters are then assigned to users through FMW security resource permissions and application policies, which then are then granted to application roles and thereby to users.

Whilst this probably seems like an over-complicated nightmare to traditional Essbase users, it does have the major advantage that one set of application roles granted to users within a Fusion Middleware system can cover both OBIEE and Essbase permissions, and there’s no need to link to Shared Services or support Native Essbase security. We’ll cover the implications of this more in some future blog posts, but this is the enabling technology that makes the rest of this integration make sense.

With Essbase integrated into the OBIEE BI Domain, you can also now use Essbase as an aggregate persistence target, though this feature comes with the same (slightly strange) approach and limitations that we first encountered when it was first introduced with OBIEE 11.1.1.6.2 BP1; although there’s not the same requirement for the Essbase server only to be used for aggregate persistence, you still have to name the Essbase database in a particular way, it’s ASO-only, and the Aggregate Persistence Wizard still creates a separate ASO database for each aggregation (similar to Oracle Database materialised views) rather than one single cube covering all aggregations. In practical terms – I’m not sure how much you’d use this vs. creating your own Essbase cube in Studio against the whole RPD business area – but it might be useful for OBIEE developers who otherwise don’t know Essbase.

So finally, the other major Essbase-related new feature in OBIEE 11.1.1.7 is SmartView, the successor to Oracle BI Office. But that’s a topic in itself, so I’ll cover that this in the next posting.

OBIEE 11.1.1.7 came out a few weeks ago, and recently we looked at Hadoop integration, one of the more interesting new features in this release. Over the next week I’ll be looking in more detail at the changes around Essbase and EPM Suite integration, but today I’ll be looking at a preview release of the upcoming OBIEE 11.1.1.7 SampleApp, provided courtesy of Philippe Lions and the BI Tech Demos team within Oracle. Philippe will be going through the new SampleApp release at the upcoming Rittman Mead BI Forum 2013 events in Atlanta and Brighton, but in the meantime lets take at what’s likely to appear in this new OBIEE 11.1.1.7 demo showcase.

As with previous OBIEE 11g SampleApps, the dashboard front page lists out all of the content, and highlights in bright blue those areas that are new to this release. The 11.1.1.7 SampleApp is largely based on earlier releases to support the 11.1.1.5 and 11.1.1.6 OBIEE versions, with any new content either showing off 11.1.1.7 new features, or adding new functional areas to the SampleApp demo.

The best place to start looking is the New Features Demo dashboard, which highlights new 11.1.1.7 features such as performance tiles, 100% stacked bar charts and waterfall charts, on this first dashboard page:

Totals within tables, pivot tables and other visualisations can now have action links associated with them, to display a financial summary report for example:

Another page on this dashboard shows of the new layout capabilities within dashboard pages; object containers can now have fixed (absolute) width and height sizes, whilst dashboard columns, and rows/columns within table and pivot table views, can also be frozen whilst other areas scroll by.

The new 11.1.1.7 SampleApp is likely to ship with Endeca Information Discovery pre-installed, and configured to provide the catalog search for OBIEE’s BI Presentation Services (a new feature in OBIEE 11.1.1.7). The SampleApp 11.1.1.7 screenshot below shows a typical “faceted search” against the web catalog, displaying key attributes and an attribute search box via an Endeca Information Discovery-style guided navigation pane. The benefit of Endeca providing catalog search vs. the Presentation Server’s regular search feature is that it looks much deeper into the catalog metadata, allows searching across many more attributes, and because it uses an in-memory index, it’s really fast.

There’s also some nice new Oracle R Enterprise content and demos, including an example where R scripts can be selected via the dashboard, parameter values supplied and the scripts then run; ORE using OBIEE’s BI Repository as a data source, and some more examples of ORE analysing the Flight Delays dataset to predict delays on future flights, for example.

11.1.17 comes with a lot of “fit and finish” tweaks to standard visualisations, and one of the dashboard pages shows of new table and pivot table features such as tooltips over row and column values, as well as features such as rollover headers, removable pivot table corners and the like – none of these are world-changing, but they’re often the sort of thing that particular customers want for their systems, and in the past we’ve had to hack-around in Javascript files and the like to meet similar requirements.

If you’re an Endeca Information Discovery developer who’s also interested in the state-of-play around OBIEE integration, there’s a whole dashboard setting out current examples around OBIEE / Endeca integration, including examples of parameter passing between OBIEE and Endeca Studio, Endeca using OBIEE’s BI Repository as its data source, and BI Publisher reporting against Endeca Server data via web service calls.

Finally, the dashboard pages added in to support DBA and developer tasks have been added to, with a new dashboard page for example displaying a list of all the physical SQL queries sent to the database.

Thanks to Philippe and the BI Tech Demos team for the early preview. Check back tomorrow when we’ll continue the look at what’s new with OBIEE 11.1.1.7, by taking a closer look at what’s changed, and dramatically improved, in the area of integration with Essbase and the Oracle EPM (Hyperion) Suite.

If you’ve not done so already, I’d advise anyone interested in OBIEE 11g to subscribe to the Oracle BI Tech Demos channel on Youtube, where Philippe Lions and his team showcase upcoming SampleApp releases, and highlight new features like SmartView integration (to be covered later this week on this blog), integration with Oracle R Enterprise, what’s coming in the new 11.1.1.7 SampleApp, and OBIEE 11.1.1.7 leveraging the Advanced Analytics Option in Oracle Database 11gR2. One demo that I’ve been aware of for some time via the Exalytics program, and that’s also featured in the Tech Demos channel, is a load test demo that uses scripts and internal OBIEE features to run the test, and is used by Oracle to show how many concurrent users an Exalytics server can handle.

What’s particularly interesting about this load test though is that it doesn’t require any external tools such as LoadRunner or JMeter, and the scripts it uses are actually shipped with the full v207 SampleApp VirtualBox VM that is downloadable from OTN. On a recent customer engagement a need came up for a “quick and dirty” load test for their system, so I thought I’d go through how this load test example works, and how it can be adapted for use with any other generic OBIEE 11g (11.1.1.6+) environment.

In the example used in the Youtube video, a report (which actually looks like a dashboard page, but is actually an single analysis compound layout containing two graph views, and a pivot table view) is set up with a special set of filter values; when requested, this analysis will use “randomised” filter values so that response times aren’t skewed by the same values being used each time, and a controlling process outside of the dashboard ramps up 10, 100, 200 and so on separate sessions up to  a maximum of 2,000, to simulate the sort of user numbers that an Exalytics server might be required to support.

Then, when the load test is running, the metric display within Fusion Middleware Control is used to show how the server copes with the load (in terms of # of sessions, average response time per query etc), as well as a dashboard page based off of the usage tracking data that shows a similar set of information.

Now of course the reason this sort of test data is important (apart from selling Exalytics servers) is that a report that takes 10 seconds to run, on a system otherwise unused and with only you running queries, might take considerably longer to run when all your users are on the system, due to factors such as disk contention, queuing on database server and mid-tier server CPUs, parallel query getting scaled-back when more than a few users try to run reports at the same time, and so on – so you need to do this sort of load test before unleashing your new dashboards onto your user community. But performing a load test is hard – just ask our Robin Moffatt – so having a ready-made system shipped with SampleApp, that doesn’t require additional software, certainly sounds interesting. So how does it work?

The scripts that control the load test process are contained within the /home/oracle/scripts/loadtest folder on SampleApp, and look like this within the Linux file manager:

The folder actually contains three scripts, and a Java JAR archive file:

• runtest actually runs the load test
• users_list.txt is a set of usernames, that are central to the load test process (more on this in a moment)
• Loadtest_README.txt is the instruction file, and
• LoadTest.jar is a Java program that is called by runtest to log into OBIEE and request the report

Looking through the readme file, the way the process works is that you need to create a set of users with a common password within the OBIEE LDAP directory, and put their usernames in the users_list.txt file. Then, the LoadTest.jar file is called by the runtest script, passing the hostname and port number of the WebLogic server hosting Presentation Services, the path to the analysis that you wish to test against, and the common password, and the script will then initiate a session for each user and then run the report.

Looking at the list of names in the users_list.txt file is interesting, because they all appear to be airport three-letter codes; for example:

SAN
SAT
SAV
SBA
SBN
SBP
MKC
MKE
MKG
MLB
MLI
MLU
MMH

The reason for this becomes clear when you look at the filters behind the analysis that the runtest script calls; to provide the filter predicate randomisation, each run of the report uses the username to filter the origin airport selection, and the other filter values are generated through MOD and RAND functions that in essence, generate random values for each call of the report. So given that we’re not all going to want to test reports based on airport codes, and how the overall testing process works, this presents two challenges to us:

1. How we generate a very large number of user accounts with a common password, given that the test process runs the report just once for each user, and how we get rid of these accounts once we’ve finished the testing.
2. How we configure the report we want to test to generate “random” filter values – the approach Oracle took with this example is quite “clever”, but we’ll need to come up with something equally clever if we want to do this for our report.

Question 1 seems extricably linked to question 2, so let’s create an example report that we can easily randomise the values for, create a number of views that we can include in a compound layout as Oracle did in the load test demo, and give it a go.

Taking the SampleApp dataset and the A – Sample Sales subject area, let’s create an analysis that has the following columns in the analysis criteria:

• Products.P2 Product Type
• Time.T03 Per Name Qtr
• Time.T02 Per Name Month
• Customer.C3 Customer Type
• Ship To Regions.R50 Region
• Base Facts.1 – Revenue
• Base Facts.1 – Discount Amount

For good measure, create another derived measure, called Base Facts.1 – Gross Revenue, which uses the formula:

• “Base Facts”.”1- Revenue”+”Base Facts”.”3- Discount Amount”

and then create some views off of this criteria so that your analysis looks something along these lines:

Now comes the tricky part of randomising it. We could take the approach that Oracle took with the Airlines load test example and create, for example, a user for each country in the dataset, but instead let’s use Logical SQL’s RAND function to pick a region and calendar quarter at random, and then three of the five customer types, to use as the analysis filters. To do this, we create a filter against this column in the analysis criteria and then convert the filter to SQL, using something like the following SQL clause to filter the quarter randomly:


"Time"."T03 Per Name Qtr" in (
SELECT a.s_1 from
(SELECT
0 s_0,
"A - Sample Sales"."Time"."T03 Per Name Qtr" s_1,
RAND()*100 s_2
FROM "A - Sample Sales"
WHERE
(BOTTOMN(RAND()*100,1) <= 1)
ORDER BY 1, 2 ASC NULLS LAST, 3 ASC NULLS LAST
FETCH FIRST 5000001 ROWS ONLY) a
)


The same goes for the region filter, which we define as:


"Ship To Regions"."R50 Region" in (
SELECT a.s_1 from
(
SELECT
0 s_0,
"A - Sample Sales"."Ship To Regions"."R50 Region" s_1,
RAND()*100 s_2
FROM "A - Sample Sales"
WHERE
(BOTTOMN(RAND()*100,1) <= 1)
ORDER BY 1, 2 ASC NULLS LAST, 3 ASC NULLS LAST
FETCH FIRST 5000001 ROWS ONLY) a )


whereas for the customer type filter, we return the top 3 ordered (random) values, not just the first one:


"A - Sample Sales"."Customers"."C3 Customer Type" in
( SELECT a.s_1
FROM
(SELECT
0 s_0,
"A - Sample Sales"."Customers"."C3 Customer Type" s_1,
DESCRIPTOR_IDOF("A - Sample Sales"."Customers"."C3 Customer Type") s_2,
RAND()*100 s_3
FROM "A - Sample Sales"
WHERE
(BOTTOMN(RAND()*100,3) <= 3)
ORDER BY 1, 2 ASC NULLS LAST, 4 ASC NULLS LAST
FETCH FIRST 5000001 ROWS ONLY) a )


Now when you run the report you should see different filter selections being used each time you run it, similar to what's shown in the preview screenshot below.

One thing I noticed at this stage is, whilst the customer type filtering returned three values, only one would ever be used in the graph prompt, because that's how prompts in a view work vs. the multi-select prompts you get as dashboard prompts. So I then needed to move the customer type column from the prompts are to the Pies and Slices > Pies part of the graph layout (so I then got one pie chart per customer type, not just the one type I was seeing via the graph prompt before), so that my final report looked like this:

and my analysis criteria, including these special filters, looked like this:

Next we need to create an initial set of users so that we can perform the concurrency test. I do this by using the WebLogic Scripting Tool (WLST) script shown below which creates 30 users, assigns them to an LDAP group and then adds that group to the  BIConsumers LDAP group, so that they can run the analysis in question (if you're new to WLST or are interested in reading a bit more about it, take a look at this Oracle Magazine of mine that explains the feature).


serverConfig()

password = 'welcome1'

atnr=cmo.getSecurityConfiguration().getDefaultRealm().lookupAuthenticationProvider('DefaultAuthenticator')

group = 'Loadtest-Users'
atnr.createGroup(group,group)

atnr.addMemberToGroup('BIConsumers','Loadtest-Users')

users = ['user1','user2','user3','user4','user5','user6','user7','user8','user9','user10','user11','user12',
'user13','user14','user15','user16','user17','user18','user19','user20','user21','user22','user23','user24',
'user25','user26','user27','user28','user29','user30']
for user in users:
atnr.createUser(user,password,user)
atnr.addMemberToGroup(group,user)


After saving the WLST script to the /home/oracle/scripts/loadtest folder as create_users.py, I then go back to my Mac workstation and SSH into the SampleApp VirtualBox VM to run the script:


Last login: Sat Apr 20 12:58:38 on ttys000
markmacbookpro:~ markrittman$ ssh oracle@obieesampleapp.rittmandev.com
oracle@obieesampleapp.rittmandev.com's password:
Last login: Sun Apr 21 17:13:37 2013 from 192.168.2.31

[oracle@obieesampleapp ~]$ cd obiee/Oracle_BI1/common/bin

[oracle@obieesampleapp bin]$ ./wlst.sh

wls:/offline> connect('weblogic','Admin123','localhost:7001')
Connecting to t3://localhost:7001 with userid weblogic ...
Successfully connected to Admin Server 'AdminServer' that belongs to domain 'bifoundation_domain'.

Warning: An insecure protocol was used to connect to the
server. To ensure on-the-wire security, the SSL port or
Admin port should be used instead.

wls:/bifoundation_domain/serverConfig> execfile('/home/oracle/scripts/loadtest/create_users.py')
Already in Config Runtime

wls:/bifoundation_domain/serverConfig> exit()

Exiting WebLogic Scripting Tool.


Then using the same SSH session I create a new users_list.txt file containing the usernames of these 30 users (use CTRL-D in a Unix session to send the EOF signal to CAT, and stop copying text into the users_list.txt.new file)


[oracle@obieesampleapp bin]$ cd /home/oracle/scripts/loadtest/

[oracle@obieesampleapp loadtest]$ cat > users_list.txt.new
user1
user2
user3
user4
user5
user6
user7
user8
user9
user10
user11
user12
user13
user14
user15
user16
user17
user18
user19
user20
user21
user22
user23
user24
user25
user26
user27
user28
user29
user30

[oracle@obieesampleapp loadtest]$ mv ./users_list.txt users_list.txt.original
[oracle@obieesampleapp loadtest]$ mv ./users_list.txt.new users_list.txt


Finally, I then edit the runtest script to change the path to point to the analysis I created earlier, update the password setting for the users:


[oracle@obieesampleapp loadtest]$ vi ./runtest

so that the final runtest file looks like this:

[oracle@obieesampleapp loadtest]$ cat ./runtest
export JAVA_HOME=/home/oracle/obiee/Oracle_BI1/jdk
export PATH=$JAVA_HOME/bin:$PATH
export CLASSPATH=$CLASSPATH:$JAVA_HOME/lib

echo "Start time: `date`"
echo "Load Test Starting..."
java -jar LoadTest.jar "localhost" "7001" "/shared/loadtest/SampleAnalysis" "welcome1"
echo "Load Test Completed..."
echo "End time: `date`"


Now, we've got everything we need for the initial test; an analysis to run, a set of users to run it with, and the JAR file to perform the test. So let's give it a go...


[oracle@obieesampleapp loadtest]$ chmod a+x runtest
[oracle@obieesampleapp loadtest]$ ./runtest
Start time: Sun Apr 21 18:21:39 PDT 2013
Load Test Starting...

----------------------------------------------
Creating User Sessions for Concurrency Test..
Total active sessions: 30

Initiating Queries..
Total queries initiated: 30

Cleaning up User Sessions created for Concurrency Test..
- Remaining Active Sessions: 30
Completed User Sessions Cleanup
----------------------------------------------

Load Test Completed...
End time: Sun Apr 21 18:21:54 PDT 2013


Where it gets interesting though is when you go over to Fusion Middleware Control, and view the DMS metrics graphs at Capacity Management > Metrics > View the full set of system metrics, where you can then see various metrics such as # of active sessions, request processing time (i.e. how long the analysis took to run), and # requests per minute.

But of course, our current test only runs thirty queries through our thirty users, so its not much of a concurrency test; also, I've got caching enabled, so I'd expect the figure to look fairly good (though this may be what we use in real-life, so the key thing is to make the test as close a reflection of your actual system as possible). 

To create a more realistic test in terms of user numbers, there's a couple of options you can use; one option, and the one I use, is to copy the same set of users over and over again into the users_list.txt file, to the point where there are hundreds of rows in the file to simulate hundreds of sessions. Another approach, and perhaps the more purist, is to create many more user accounts and have each one only run one or two reports, but that involves creating the required amount of users and then deleting them afterwards from the LDAP server. I'll go for the first option, using the following Unix commands within my SSH session to copy the file back onto itself many times, giving me at the end around 1600 users to use in my concurrency test:


[oracle@obieesampleapp loadtest]$ wc -l users_list.txt
30 users_list.txt
[oracle@obieesampleapp loadtest]$ cat users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt users_list.txt >> big_users_list.txt
[oracle@obieesampleapp loadtest]$ wc -l big_users_list.txt
420 big_users_list.txt
[oracle@obieesampleapp loadtest]$ cat big_users_list.txt big_users_list.txt big_users_list.txt big_users_list.txt > users_list.txt
[oracle@obieesampleapp loadtest]$ wc -l users_list.txt
1680 users_list.txt


FInally I run the test again, to simulate 1600 users running queries at once:


[oracle@obieesampleapp loadtest]$ ./runtest
Start time: Sun Apr 21 18:42:42 PDT 2013
Load Test Starting...

----------------------------------------------
Creating User Sessions for Concurrency Test..
- Active Sessions: 100
- Active Sessions: 200
- Active Sessions: 300
- Active Sessions: 400
- Active Sessions: 500
- Active Sessions: 600
- Active Sessions: 700
- Active Sessions: 800
- Active Sessions: 900
- Active Sessions: 1000
- Active Sessions: 1100
- Active Sessions: 1200
- Active Sessions: 1300
- Active Sessions: 1400
- Active Sessions: 1500
- Active Sessions: 1600
Total active sessions: 1680

Initiating Queries..
- Queries initiated: 100
- Queries initiated: 200
- Queries initiated: 300
- Queries initiated: 400
- Queries initiated: 500
- Queries initiated: 600
- Queries initiated: 700
- Queries initiated: 800
- Queries initiated: 900
- Queries initiated: 1000
- Queries initiated: 1100
- Queries initiated: 1200
- Queries initiated: 1300
- Queries initiated: 1400
- Queries initiated: 1500
- Queries initiated: 1600
Total queries initiated: 1680

Cleaning up User Sessions created for Concurrency Test..
- Remaining Active Sessions: 1680
- Remaining Active Sessions: 1600
- Remaining Active Sessions: 1500
- Remaining Active Sessions: 1400
- Remaining Active Sessions: 1300
- Remaining Active Sessions: 1200
- Remaining Active Sessions: 1100
- Remaining Active Sessions: 1000
- Remaining Active Sessions: 900
- Remaining Active Sessions: 800
- Remaining Active Sessions: 700
- Remaining Active Sessions: 600
- Remaining Active Sessions: 500
- Remaining Active Sessions: 400
- Remaining Active Sessions: 300
- Remaining Active Sessions: 200
- Remaining Active Sessions: 100
Completed User Sessions Cleanup
----------------------------------------------

Load Test Completed...
End time: Sun Apr 21 18:45:34 PDT 2013


Going back over to EM, I can see the load building up on the server and the response time increasing.

Notice though how the response time actually starts to fall as more queries run? That's most probably caching kicking in, so next time I'll disable caching completely and run the test again. But for now though, this is the Oracle load test script running, and the steps I've outlined here should allow you to run a similar test yourself. Thanks to Phillipe and the Oracle BI Tech Demos team for this, and on a similar topic I'll be previewing the new v303 11.1.1.7 SampleApp in a posting tomorrow.

A few days back, I introduced our special guests for the Rittman Mead BI Forum in Atlanta, focusing first on Cary Millsap. Today I’d like to talk about our other special guest: Oracle ACE Director Alex Gorbachev. Alex was an inspiration for me back in the Oracle Database 10g and early 11g days when I was administering Oracle RAC for several data warehouse customers, and wondering whether RAC was the right platform for BI. Of course it was… and every time I read one of Alex’s blogs (he was quite a prolific blogger back then… we all were once upon a time) or saw him speak, I felt empowered to go take on Cache Fusion.

Alex joined Pythian in Canada as a DBA team lead in 2006. Just two years later, he moved to Australia to successfully startup Pythian Australia. In 2009, he returned to Canada and took up the mantle of Chief Technology Officer, a title he still holds today. He is a member of the distinguished OakTable Network (as is Cary Millsap… something I forgot to mention yesterday), and is a member of the Board of Directors of the Independent Oracle Users Group (IOUG). Alex founded the Battle Against Any Guess Party, a movement promoting scientific troubleshooting techniques. During his time in Australia he also founded Sydney Oracle Meetup, a vibrant local community of passionate Oracle professionals.

Its fortuitous that Mark blogged yesterday on Hadoop… as this is exactly what Alex is speaking on at the BI Forum. His presentation is titled “Hadoop versus the Relational Data Warehouse.” He’ll discuss some of the technical design principles of Hadoop and the reasons for it’s rise in popularity. We’ll get to see the position that Hadoop currently occupies in the enterprise data center, it’s possible future trajectory, and how that trajectory compares with the more traditional relational data warehouse. For the BI developers in the crowd who have perhaps never seen Alex speak… you’re definitely in for a treat. He’s set to speak first thing Friday morning to kick off the last day of the Forum. If you know Alex, you’re obviously aware that he’s an excellent technologist, but you also likely know how much fun he is to be around, so it will be good to have him at the social meet-ups in and around the conference.

I’d really like to thank our friend and business partner Pythian for always supporting Rittman Mead and ensuring that Alex would speak at the Forum. And of course… I’d be remiss if I didn’t say: Love Your Data!

In yesterday’s post I looked at the key enabling technologies behind OBIEE and ODI’s connectivity to Hadoop, and today I’ll look at how OBIEE 11.1.1.7 can now access Hadoop data sources through two related technologies; Hive, and MapReduce.

In my introduction to the topic I said that whilst writing MapReduce routines in Java, and then orchestrating them through other tools in the Apache Hadoop family could be quite complex technically, another tool called “Hive” provided an SQL-like query layer over Hadoop and MapReduce so that tools like OBIEE could access them. Rather than you having to write your own MapReduce routines in Java, for instance, Hive writes them for you, returning data to OBIEE and ODI via ODBC and JDBC drivers. The diagram below, also from yesterday’s post, shows the data layers used in such an arrangement.

Under the covers, Hive has its own metadata layer, server engine and data store, with developers “loading” data into Hive “tables” which are then generally stored on the HDFS file system, just like any other data processed through MapReduce. Then, when a query is issued through Hive, the Hive Server dynamically generates MapReduce routines to query the underlying data, returning data to users in a similar way to an interactive database SQL session, like this:


markmacbookpro:~ markrittman$ ssh oracle@bigdatalite
oracle@bigdatalite's password:
Last login: Wed Apr 17 04:02:59 2013 from 192.168.2.200
=====================================================
=====================================================
Welcome to BigDataLite
run startx at the command line for X-Windows console
=====================================================
=====================================================

Host: bigdatalite.us.oracle.com [192.168.2.35]

[oracle@bigdatalite ~]$ hive
Hive history file=/tmp/oracle/hive_job_log_oracle_201304170403_1991392312.txt

hive> show tables;
OK
dwh_customer
dwh_customer_tmp
i_dwh_customer
ratings
src_customer
src_sales_person
weblog
weblog_preprocessed
weblog_sessionized
Time taken: 2.925 seconds

hive> select count(*) from src_customer;
Total MapReduce jobs = 1
Launching Job 1 out of 1
Number of reduce tasks determined at compile time: 1
In order to change the average load for a reducer (in bytes):
set hive.exec.reducers.bytes.per.reducer=
In order to limit the maximum number of reducers:
set hive.exec.reducers.max=
In order to set a constant number of reducers:
set mapred.reduce.tasks=
Starting Job = job_201303171815_0003, Tracking URL = http://localhost.localdomain:50030/jobdetails.jsp?jobid=job_201303171815_0003
Kill Command = /usr/lib/hadoop-0.20/bin/hadoop job -Dmapred.job.tracker=localhost.localdomain:8021 -kill job_201303171815_0003
2013-04-17 04:06:59,867 Stage-1 map = 0%, reduce = 0%
2013-04-17 04:07:03,926 Stage-1 map = 100%, reduce = 0%
2013-04-17 04:07:14,040 Stage-1 map = 100%, reduce = 33%
2013-04-17 04:07:15,049 Stage-1 map = 100%, reduce = 100%
Ended Job = job_201303171815_0003
OK
25
Time taken: 22.21 seconds

hive>


In the example above, I connected to the Hive environment, listed out the “tables” available to me, and then ran a count of “rows” in the src_customers table which in the background, caused a MapReduce routine to be written and executed in the background by the Hive server. Hive has been described as the “Hadoop Data Warehouse”, but it’s not really a data warehouse as you and I would know it – you wouldn’t typically use Hadoop and Hive to store customer transaction data, for example, but you might use it as a store of Facebook interactions, for example, or most popular pages or interaction paths through your website, and someone working in Web Analytics might want to interactively query that set of data in a more user-friendly manner than writing their own Java routines. So how does OBIEE gain access to this data, and what extra software or configuration pieces do you need to put in-place to make it happen?

If you want to have OBIEE 11g access Hadoop data, you’re best going with the 11.1.1.7+ release as this is where it’s most tested and stable. You’ll need to configure drivers at two points; firstly at the server level (Hadoop access is only supported with Linux server installations of OBIEE 11.1.1.7) and then at the Windows-based Administration tool level. Let’s start with the BI Administration tool first, based on the instructions in the 11.1.1.7 Metadata Repository Builder’s Guide.

To have the BI Administration tool connect to a Hadoop/Hive data source, you’ll need to download some ODBC drivers for Hadoop via a My Oracle Support download, DocID 1520733.1. This gives you a set of HiveODBC drivers along with a PDF explaining the installation process, and once you’ve installed the drivers, you’ll need to open up the ODBC Data Source Administrator applet and create a new HiveODBC data source. In this instance, I call the datasource “bihdatalite” after the server name, and go with the default values for the other settings. Note that “default” is the name of the “database” within Hive, and the port number is the port that the Hive server is running on.

Now I can create a new repository offline, and connect to the Hive server via the HiveODBC connection to start importing table metadata into the RPD. Note that with the current implementation of this connectivity, whilst you can import tables from multiple Hive databases into the RPD, queries you issue can’t span more than a single Hive database (i.e. you can’t specify a schema name prefix for the table name, therefore can’t join across two schemas).

Then, once you’ve imported the Hive table metadata into the RPD, change the physical database type to “Apache Hadoop”, from the default ODBC 3.5 setting that would have been added automatically by the metadata import process. Leave the connection pool call interface at ODBC2.0, put in any old username and password into the shared login details (or a valid username/password if Hive security is enabled), and then save the repository.

You should then be able to use the View Data feature in the BI Administration tool to view data in a particular Hive table, like this:

Now you need to move over to the server part of OBIEE, and configure the ODBC connection to Hive there too. OBIEE 11.1.1.7 comes with DataDirect drivers already installed that will connect to Hive, so it’s just a case then of configuring a connection of the same name to the Hive datasource using OBIEE’s odbi.ini file, like this:


[ODBC Data Sources]
AnalyticsWeb=Oracle BI Server
Cluster=Oracle BI Server
SSL_Sample=Oracle BI Server
bigdatalite=Oracle 7.1 Apache Hive Wire Protocol

[bigdatalite]
Driver=/u01/app/Middleware/Oracle_BI1/common/ODBC/Merant/7.0.1/lib/ARhive27.so
Description=Oracle 7.1 Apache Hive Wire Protocol
ArraySize=16384
Database=default
DefaultLongDataBuffLen=1024
EnableLongDataBuffLen=1024
EnableDescribeParam=0
Hostname=bigdatalite
LoginTimeout=30
MaxVarcharSize=2000
PortNumber=10000
RemoveColumnQualifiers=0
StringDescribeType=12
TransactionMode=0
UseCurrentSchema=0


Note that you also need to configure OBIEE’s OPMN feature to use the DataDirect 7.1 drivers rather than the default, older ones – see the docs for full details on this step. Then, as far as the RPD is concerned, you just need to make a business model out of the Hive table sources, upload it using EM so that its running online on your OBIEE server installation, and your RPD in the end should look similar to this:

Then finally, you can create an OBIEE analysis using this data, and analyse it just like any other data source – except, of course, that there’s quite a lot of lag and latency at the start of the query, as Hive spins up its Java environment, writes the MapReduce query, and then send the data back to OBIEE’s BI Server.

So how do we get data into Hive in the first place, to create these tables that in the background, are access through Hadoop and MapReduce? Check back tomorrow, when I’ll look at how Oracle Data Integrator can be used to load data into Hive, as well as perform other data integration tasks using Hadoop and other big data technologies.

I always find myself putting an asterisk (if only mentally) next to certain statements. I shall now put all those statements here and link back.

1. I don't know everything
2. I'm not the best developer in the world, but I constantly work at getting better...
3. If I make a statement about something, that's been my experience. Your results may vary.

That's it...for now.

Recent releases of OBIEE and ODI have included support for Apache Hadoop as a data source, probably the most well-recognised technology within the “big data” movement. Most OBIEE and ODI developers have probably heard of Hadoop and MapReduce, a data-processing programming model that goes hand-in-hand with Hadoop, but haven’t tried it themselves or really found a pressing reason to use them. So over this next series of three articles, we’ll take a look at what these two technologies actually are, and then see how OBIEE 11g, and also ODI 11g connect to them and make use of their features.

Hadoop is actually a family of open-source tools sponsored by the Apache foundation that provides a distributed, reliable shared storage and analysis system. Designed around clusters of commodity servers (which may actually be virtual and cloud-based) and with data stored on the server, not on separate storage units, Hadoop came from the world of Silicon Valley social and search companies and has spawned a raft of Apache foundation sub-projects such as Hive (for SQL-like querying of Hadoop clusters), HBase (a distributed, column-store database based on Google’s “BigTable” technology), Pig (a procedural language for writing Hadoop analysis jobs that’s PL/SQL to Hive’s SQL) and HDFS (a distributed, fault-tolerant filesystem). Hadoop, being open-source, can be downloaded for free and run easily on most Unix-based PCs and servers, and also on Windows with a bit of mucking-around to create a Unix-like environment; the code from Hadoop has been extended and to an extent commercialised by companies such as Cloudera (who provide the Hadoop infrastructure for Oracle’s Big Data Appliance) and Hortonworks, who can be though of as the “Red Hat” and “SuSE” of the Hadoop world.

MapReduce, on the other hand, is a programming model or algorithm for processing data, typically in parallel. MapReduce jobs can be written, theoretically, in any language as long as they exposure two particular methods, steps or functions to the calling program (typically, the Hadoop Jobtracker):

• A “Map” function, that takes input data in the form of key/value pairs and extracts the data that you’re interested in, outputting it again in the form of key/value pairs
• A “Reduce” function, which typically sorts and groups the “mapped” key/value pairs, and then typically passes the results down to the line to another MapReduce job for further processing

Joel Spolsky (of Joel on Software fame, one of mine and Jon’s inspirations in setting up Rittman Mead) explains MapReduce well in this article back from 2006, when he’s trying to explain the fundamental differences between object-orientated languages like Java, and functional languages like Lisp and Haskell. Ironically, most MapReduce functions you see these days are actually written in Java, but it’s MapReduce’s intrinsic simplicity, and the way that Hadoop abstracts away the process of running individual map and reduce functions on lots of different servers , and the Hadoop job co-ordination tools take care of making sense of all the chaos and returning a result in the end, that make it take off so well and allow data analysis tasks to scale beyond the limits of just a single server..

I don’t intend to try and explain the full details of Hadoop in this blog post though, and in reality most OBIEE and ODI developers won’t need to know how Hadoop works under the covers; what they will often want to be able to do though is connect to a Hadoop cluster and make use of the data it contains, and its data processing capabilities, either to report against directly or more likely, use as an input into a more traditional data warehouse. An organisation might store terabytes or petabytes of web log data, details of user interactions with a web-based service, or other e-commerce-type information in an HDFS clustered, distributed fault-tolerant file system, and while they might then be more than happy to process and analyse the data entirely using Hadoop-style data analysis tools, they might also want to load some of the nuggets of information derived from that data in a more traditional, Oracle-style data warehouse, or indeed make it available to less technical end-users more used to writing queries in SQL or using tools such as OBIEE.

Of course, the obvious disconnect here is that distributed computing, fault-tolerant clusters and MapReduce routines written in Java can get really “technical”, more technical than someone like myself generally gets involved in and certainly more technical than you average web analytics person will want to get. Because of this need to provide big-data style analytics to non-Java programmers, some developers at Facebook a few years ago came up with the idea of “Hive”, a set of technologies that provided a SQL-type interface over Hadoop and MapReduce, along with supporting technologies such as a metadata layer that’s not unlike the RPD that OBIEE uses, so that non-programmers could indirectly create MapReduce routines that queried data via Hadoop but with Hive actually creating the MapReduce routines for you. And for bonus points, because the HiveQL language that Hive provided was so like SQL, and because Hive also provided ODBC and JDBC drivers conforming to common standards, tools such as OBIEE and ODI can now access Hadoop/MapReduce data sources and analyse their data just like any other data source (more or less…)

So where this leaves us is that the 11.1.1.7 release of OBIEE can access Hadoop/MapReduce sources via a HiveODBC driver, whilst ODI 11.1.1.6+ can access the same sources via a HiveJDBC driver. There is of course the additional question as to why you might want to do this, but we’ll cover how OBIEE and then ODI can access Hadoop/MapReduce data sources in the next two articles in this series, as well as try and answer the question as to why you’d want to do this, and what benefits OBIEE and ODI might provide over more “native” or low-level big data query and analysis tools such as Cloudera’s Impala or Google’s Dremel (for data analysis) or Hadoop technologies such as Pig or Sqoop (for data loading and processing). Check back tomorrow for the next instalment in the series.

In the previous two posts in this series, I looked at the product architecture for Oracle Enterprise Manager 12cR2 (EM12cR2) Cloud Control and the BI Management Pack, and how you registered OBIEE, TimesTen, Essbase and the DAC as targets for monitoring and managing. But what can you do with EM12cR2 and the BI Management Pack once you’ve set it all up, how well does it handle other related products such as Informatica and Siebel CRM, how customisable is it and what other tasks can it perform?

To start off, one of the questions we’ve been asked is whether, in a similar way to OBIEE and Oracle Portal, you can customise the EM web console display to include just those views that you’re interested in; to create, for example, a dashboard page for monitoring OBIEE that might include views on BI Server query throughput, GoldenGate activity, DAC ETL alerts and so on. The answer is – not quite – but there are some customisations and bookmarks that you can create which at least make it easier to navigate your way around. 

When you first log into OEM12cR2, you’re presented with the standard Enterprise Summary view, which summarises a number of metrics across all targets in the EM repository.

You can, however, change this for a more focused view of a particular type of target, by selecting SYSMAN > Select My Home… (or whatever your logged-in user name is), and then selecting from the list of pre-defined target views presented on the Select Enterprise Manager Home page that’s then displayed.

If, for example, your primary responsibility was looking after OBIEE systems, you might choose to have the Middleware page as your homepage, so that all of the WebLogic farms are listed out on your front page.

You can also set individual pages in EM as “favorites”, so that they appear from the Favorites menu for quick access as shown in the screenshot below.

Something else that’s useful when you’ve got a number of similarly-named systems registered within your EM repository is to put them into groups. To create a group to hold my “demo” OBIEE systems, for example, I would select Targets > Groups from the web console menu, and then press the Create > Group button to bring up the Create Group page. Then, using the Search or Search by Criteria buttons I can refine the search to include, for example, just Fusion Middleware Farms, and then select the ones that I’d like to add to the new group.

You can also create “dynamic” groups as well, including all systems that have a “development” status in a group that updates over time, like this:

Once you’ve registered your systems, you can do all of the same things you did with the 10gR4 version of EM and the BI Management Pack, including view metrics over time rather than for just the time you’ve got the metric window open (to my mind, one of the most valuable features in EM vs. Fusion Middleware Control).

Metric thresholds can also be defined in a similar fashion to how they were in EM10gR4, with events that are then triggered by the threshold being exceeded to notify you, for example, when query response times exceed a certain number of seconds, or when the dashboard login page can’t be reached. Unfortunately the dashboard and scheduler reports that are included as part of the BI Management Pack can’t be turned into graphs, but like Fusion Middleware Control any of the standard metrics can be graphed, overlayed on the same server’s metrics for the previous day, or compared to another server’s metrics or a baseline.

Finally, another question we’re often asked is how many other systems EM12cR2 can monitor, either out-of-the-box, through paid-for official plugins, or through third party extensions? The first thing to be aware of then is what EM functionality is included “for free” as part of your database or middleware license and what functionality costs more, and the definitive place for this information is the Oracle® Enterprise Manager Licensing Information 12c Release 2 (12.1.0.2) doc on OTN; also from the web console you can select Setup > Management Packs > Show Management Pack Information to have EM highlight for you those menu items that require additional licensing beyond those included by default for licensed database or middleware customers. For example, in the example below the items annotated with “OBIM” would require an Oracle BI EE customer to purchase the BI Management Pack, whilst the others would be “free” to use by any BI customer.

As for what these management pack and plug-ins cost, again the definitive source is the Oracle Tech Price list, which changes from time to time but can always be found with a Google search for “oracle tech price list”. The price list as of the time of writing listed the BI Management Pack at $11,500/processor (based on the processors licensed for BI EE).

Note also with management packs that you generally – at least in the case of Oracle Database – need to license the appropriate database option as well, though plug-ins are generally free or at least provided as part of the main product cost, as is the case with TimesTen and Exadata. In terms of what features come out of the box and what ones require separate installation, you can check this by selecting Setup > Extensibility > Self Update and Plugins menu items, which show the downloaded and available agent versions, along with the various plugins that can be used immediately, or downloaded from Oracle’s support site, including ones for Siebel, below, and EMC’s SAN arrays.

There are also plug-ins available for download from third-party sites for targets such as ones for Informatica PowerCenter, VMWare VSphere and mySQL, with most of them gathered together at the Enterprise Manager Extensions Exchange, also on the Oracle website.

So there we are with our three-part look at EM12cR2 and the BI Management Pack. I’m over in Norway now for the Oracle User Group Norway conference, but check back soon for some new content on the 11.1.1.7 release of OBIEE 11g.

I feel like I’m introducing the Beatles… though I think Kellyn Pot’Vin calls them the “DBA Gods”.  Today I’ll be talking about Cary Millsap, and tomorrow I’ll introduce our other special guest: Alex Gorbachev.

As many of you know, I grew up as a DBA (albeit, focusing on data warehouse environments) before transitioning to development… initially as an ETL developer and later as an OBIEE architect. I had three or four “heroes” during that time… and Cary Millsap was certainly one of them. His brilliant white paper “Why a 99%+ Database Buffer Cache Hit Ratio is Not Ok” changed my whole direction with performance tuning: it’s probably the first time I thought about tuning processes instead of systems. Many of you also know about my love of Agile Methodologies… a cause that Cary has championed of late, and is also the subject of an excellent white paper “Measure Once, Cut Twice”. This purposeful inversion of the title helps to remind us that many of the analogies we use for software design don’t compute… it’s relatively simple to modify an API after the fact, so go ahead and “cut”.

A brief bit of history on Cary. He’s an Oracle ACE Director and has been contributing to the Oracle Community since 1989. He is an entrepreneur, software technology advisor, software developer, and Oracle software performance specialist. His technical papers are quoted in many Oracle books, in Wikipedia, in blogs all over the world, and in dozens of conference presentations each month. He has presented at hundreds of public and private events around the world, and he is published in Communications of the ACM. He wrote the book “Optimizing Oracle Performance” (O’Reilly 2003), for which he and co-author Jeff Holt were named Oracle Magazine’s 2004 Authors of the Year. Though many people (Kellyn included) think of Cary as a DBA… Cary considers himself to be software developer first, but explains what he believes to be the reason for this misconception:

“I think it’s fair to say that I’ve dedicated my entire professional career (27 years so far) to the software performance specialization. Most people who know me in an Oracle context probably think I’m a DBA, because I’ve spent so much time working with DBAs (…It’s still bizarre to me that performance is considered primarily a post-implementation operational topic in the Oracle universe). But my background is software development, and that’s where my heart is. I built the business I own so that I can hang out with extraordinarily talented software researchers and designers and developers and write software that helps people solve difficult problems.”

Cary’s presentation is called “Thinking Clearly about Performance” and will be given at the end of the day on Thursday before we head over to 4th and Swift for the Gala dinner. His message for the BI developers in the audience is an encouraging one:

“My message at the Rittman Mead BI Forum is that, though it’s often counterintuitive, software performance actually makes sense. When you can think clearly about it, you generally make progress more quickly and more permanently than if you just stab at different possible solutions without really understanding what’s going on. That’s what this presentation called “Thinking Clearly about Performance” is all about. It’s the result of more than 25 years of helping people understand performance in just about every imaginable context, from the tiny little law office in east Texas to the Large Hadron Collider at CERN near Geneva. It’s the result of seeing the same kinds of wasteful mistakes: buying hardware upgrades in hopes of reducing response times without understanding what a bottleneck is, adding load to overloaded systems in hopes of increasing throughput.”

My experience is that BI and DW systems suffer more from the “fast=true” disease than do OLTP systems, but that could simply be perspective. I’m excited that a group of BI developers, the majority of which are reporting against a database of some kind, will get an opportunity to see Cary’s approach to problem solving and performance tuning. As Cary tells us:

“The fundamental problems in an OBIEE implementation are just that: fundamental. The solution begins with understanding what’s really going on, which means engaging in a discussion of what we should be measuring and how (of course, in the OBIEE world, Robin Moffatt’s blog posts come in handy), and it continues through the standard subjects of profiles, and skew, and efficiency, and load, and queueing, and so on.”

If you are interested in seeing Cary and all the other great speakers at this year’s BI Forum, you can go over to the main page to get more information about the Atlanta event, or go directly to the registration page so we can see you in Atlanta in May.

In the previous post in this series, we looked at what Oracle Enterprise Manager 12cR2 and the BI Management Pack could do for OBIEE 11g admins, and how it manages a number of Oracle products from the database through to Fusion Middleware and the ERP applications. In today’s post I’m going to look at how an OBIEE system (or “target”) is registered so that we can then use BI Management Pack features, and how you make use of new features in the BI Management Pack such as support for Essbase targets.

I’ll work on the assumption that you’ve already got EM12cR2 installed, either on 64-bit Windows or 64-bit Unix (my preference is 64-bit Oracle Linux 5, though all should work); if you’ve not got EM12cR2 installed or your on an earlier version, the software is available on OTN and you’ll also need a suitable, recent and patched-up database to store the EM repository. Once you’ve got everything installed we can now login and take a look around – note that there’s no separate BI Management Pack download; all functionality is included but you need to be aware of what’s extra-cost and what’s not – the licensing guide is your best reference here, but at a high-level there are some parts of EM12cR2 that all licensed BI customers can use, whilst other features require the BI Management pack – we’ll cover this in more details in tomorrow’s post.

Logging into EM12cR2 presents you with a summary of the health and status of your systems, and you can see from the pie chart on the left that some of my systems are up, some are down and so forth. The Targets menu at the top of the screen lets me view similar information for hosts, middleware installations, databases and so on. My EM12cR2 installation has a number of OBIEE and other systems already registered with it, all of which are on VMs of which only a few are currently powered up.

In this example, I’m going to add a new host to this list, which is actually an Exalytics demo VM containing OBIEE 11.1.1.6 and TimesTen. Later on, we’ll look at adding Essbase to the list of monitored targets, both in terms of Essbase integrated into an OBIEE 11.1.1.7 install, and standalone as a separate install; finally, we’ll see how the BI Apps DAC is registered, so we can view the progress of Informatica ETL runs into the BI Apps data warehouse.

As I mentioned in yesterday’s post, EM12cR2 monitors and manages other servers by installing management agents on them; to do this, it needs to connect to the server via SSH, in order to install the agents software on there. To enable this, you need to provide the login credentials for a user on that server with “sudo” (act as an administrator) privileges, and a number of other settings have to be enabled for this process to work; to check that all of these are in place, let’s open up a console session on the Exalytics server and see how it looks:


[oracle@exalytics ~]$
[oracle@exalytics ~]$ sudo vi /etc/hosts
[sudo] password for oracle:
oracle is not in the shudders file. This incident will be reported.
[oracle@exalytics ~]


What happened here is that I tried to run a command as the superuser, and the system asked for my password, but it turns out that this user isn’t in the list of users authorised to act as the superuser. To fix this, I need to now actually log in as root, and then issue the command:

/usr/sbin/visudo

to open up a special version of “vi” used for editing the sudoers file, and then add the line:


oracle ALL=(ALL) ALL


to the end to enable the “oracle” user to use the sudo command. After doing this and trying the previous commands again, I can now use the sudo command. Let’s now move over to the EM12cR2 website and start the process of registering the host, and thereafter registering the various software components on the Exalytics server.

There are various automated and semi-automated ways of discovering candidate servers on your network, but for simplicity I just select Setup > Add Target > Add Targets Manually from the menu in to top right-hand corner of the screen, which allows me to add details of the host directly rather than let EM scan the network to find them.

The Add Targets Manually page is then displayed. I select Add Host Targets from the set of radio button options, and then press the Add Host … button.

I then type in the name of the host (or IP address), and select the platform type. Note that unless your target is Linux x64, you’ll probably need to download the required agent software for that platform before you perform this task, using the Self Update feature.

Then, type in the location on the remote server that you want to install the agent software to, and the login credentials of the user that you’ve just enabled for sudo access on that server.

EM12cR2 will then attempt to install the agent. If you’re lucky, it’ll install first time, but more likely you’ll see a message like the one in the screenshot below saying that there was a problem with the agent installation.

What this is saying is that you need to make some further changes to the “sudoers” file on the remote server before EM can properly use it to install the agent. There are usually actually two issues, and you hit the second one after fixing the first, so let’s tackle them both now. Going back over to the remote server and logging in as the “oracle” user, let’s use sudo again to fix the issue:


[oracle@exalytics ~]$ sudo /usr/sbin/visudo


 The first bit to find in the file is the line that says (assuming Oracle Linux 5, as I’m using):

Defaults  requiretty

This setting means that all users trying to use the sudo command need to be actually logging in via the server’s own console, not remotely via SSH; by default this is the setting with Oracle Linux 5, so to change it to not require access through the console for users, I change it to:

Defaults   !requiretty

While you’re there, also add the line:

Defaults   visiblepw

to the file as well, as EM will complain about that not being set if you try and deploy again now. Once both of these are set, go back to EM, retry the deployment with the existing settings, and the agent should deploy successfully. Note also that if your OBEE installation is on a Windows server, you’ll need to install the cygwin Unix-like environment on the server before you do all this, to enable the SSH service and BASH command shell that EM requires – see these notes in the EM12cR2 docs for more details.

So at this point the management agent software will be deployed to the server, but none of the WebLogic, database or BI software will be registered with EM yet. To do this, on the screen that’s again displayed after you’ve registered the host itself, select Add Non-Host Targets using Guided Process (Also Adds Related Targets) option, then from the Target Type drop-down menu select Oracle Fusion Middleware, and then press the Add Using Guided Discovery… button to start the process by registering the WebLogic installation which in turn hosts OBIEE 11g.

When prompted, select the host that you’ve just installed the agent to as the WebLogic Administration Server Host, put in the web logic administration user details to connect to the admin server (not the OS user details you used earlier), leave the rest of the settings at their default values and press Continue.

If all goes to plan EM should then report a number of targets found in the scan – these are the WebLogic Server components, plus the BI components that we’re actually interested in.

On the next page, add any notes that you want to the target registration details, then press the Add Targets button add these to the EM repository.

On the Middleware targets page that is displayed once the targets are registered, you should see your WebLogic installation now listed, and if you drill into the Farm entry you’ll see the domain and the coreapplication entry that represents your Oracle instance. Click on the details for the farm, and you’ll then see something that looks familiar from Fusion Middleware Control – the view of your OBIEE installation, where you can also drill into core application and see details of your instance. We’ll cover more on what this screen can do in tomorrows, final post on this topic.

At this point our OBIEE system is mostly registered and configured, but we still need to register the repository database, so the dashboard and scheduler reports can work. To do this, select Business Intelligence Instance > Target Setup > Monitoring Credentials from the coreapplication drop-down menu, and then enter the details for that server’s BIPLATFORM schema, like this:

You should then be able to select Business Intelligence Instance > Dashboard Reports from the coreapplication drop-down menu to see details of which dashboards have run, what error messages were logged and so forth.

Note that this is a fairly minimal set of reports against usage tracking data – there’s no ability to graph the results, for example, and no ability to view individual report usage, just dashboards. But at least it’s something.

So that’s taken care of the OBIEE elements of the Exalytics server install. But what about TimesTen server that provides the in-memory database cache on the Exalytics server? TimesTen support doesn’t come out-of-the-box with EM12cR2, but you can enable it through a plug-in that you enable via EM12cR2′s “self-update” feature. To do this, from the Setup menu on the top-right hand side select Setup > Extensibility > Self Update, click on Plug-in in the list of folders that are then displayed, and then locate the Oracle TimesTen In-Memory Database entry in the Plug-in Updates listing that is then displayed. Assuming that its not been downloaded by someone else beforehand, click on it and press the Download button, to start the download process into EM’s software library.

After a short while the plug-in should be downloaded from Oracle’s support website, and you can then make it available for use with an agent. To do so, locate it in the list of agents again, click on it to select it, and then press the Apply button that’s next to the (greyed-out) Download button. You’ll then be taken to the Plug-ins page where you should use the Deploy On button to deploy it first to the Management Server (i.e. the EM12cR2 server) and then the Management Agent server (in this case, our Exalytics server) – note that you’ll need to know the SYS password for the database that holds your EM repository to do the OMS registration part.

If all goes to plan EM should then start the process of deploying the TimesTen plugin to the Management Server first, once its checked prerequisites and so forth. On my system, it also actually deployed the plug-in to the Exalytics server too, even though I don’t think I actually requested it.

The final configuration step now is to use the plug-in to register the TimesTen target on the Exalytics server. To do this I return to the main Setup menu in the EM web console, and select Setup > Add Target > Add Targets Manually, and then select the Add Non-Host Targets by Specifying Target Monitoring Properties radio button. Then, when the Target Type drop-down menu is displayed, select TimesTen In Memory Database 11g from the list, and then select the management agent that’s on the Exalytics server. Once done, press the Add Manually… to go on to the next stage of the target registration process.

Then, when prompted, enter the connection details to your TimesTen instance, as used on the Exalytics server.

And that should be it – the TimesTen server should be registered and then available as a target to view in EM. It’ll take a while for metrics to start getting collected and displayed in the various graphs, but you can take a look at what’s recorded and what actions you can take from the menu that’ll now appear when you view a TimesTen target.

For Essbase, how to register the Essbase server as a target depends on whether Essbase is installed standalone in just a WebLogic managed server (as it is with Oracle’s SampleApp demo VMs), installed alongside OBIEE 11.1.1.7 or 11.1.1.6.2 BP1 as part of a single BI domain, or installed in its own full WebLogic domain with a WebLogic Server administration server. If its installed standalone, the initial registration of the WebLogic domain on the server concerned won’t register the Essbase server, and instead you’ll need to register it manually afterwards in a similar manner to the TimesTen server. If you’ve installed Essbase along with OBIEE as part of the combined 11.1.1.7 install, it’ll get registered along with OBIEE, and be displayed underneath coreapplication as shown in the screenshot below. Finally, if Essbase has its own WebLogic domain, then it gets detected as a target type as part of that domain’s registration, the same way that OBIEE does when registering it’s WebLogic domain as a target.

Finally, the BI Apps Data Warehouse Administration Console (DAC) is registered similarly to Essbase and TimesTen, except that like Essbase the plug-in required for management is already included in most EM12cR2 installations. As the DAC isn’t associated with any particular middleware home (at least, not with BI Apps 7.9.6.3) you’ll need to find it within the general list of targets rather than with the OBIEE installation its linked with.

So, with all of this set up, what can you do with it? In the final part of this series tomorrow, we’ll look at some common questions and usage scenarios around EM12cR2 and the BI Management Pack, and try and answer some of these questions.

A few years ago I wrote a series of blog posts, and an OTN article, on managing OBIEE 10g using Oracle Enterprise Manager 10gR4 and the BI Management Pack, an extra-licensable option for OEM that provided additional management capabilities for OBIEE  and the BI Apps Data Warehouse Administration Console. The BI Management Pack was reasonably popular at the time but disappeared with the move to Enterprise Manager 12c Cloud Control, but with the recent release of EM12cR2 it’s come back again, but now with additional capabilities around WebLogic, GoldenGate, TimesTen and Essbase. I covered the news of this new release a few months ago, and since then our customers are often asking about these new capabilities, but information on Oracle’s website and the web is pretty thin so I thought I’d go through it in a bit more detail, today talking about how the product works, tomorrow going through installation and configuration and then on the third day, covering some of the common requests and questions we’ve had from our own customers.

Unlike Oracle Enterprise Manager Fusion Middleware Control (or indeed Database Control, the equivalent for the Oracle Database), Enterprise Manager 12cR2 Cloud Control is designed to manage multiple target systems, not just the one that its installed on. What this means is that you can manage all of your BI domains from one place, along with all of your databases, your GoldenGate installation, the DAC, Essbase and so forth, with their details held in a management repository stored in an Oracle database. The diagram below shows a typical OEM12cR2 topology, with the OEM installation on a server connected to the repository database, and OBIEE and other BI “targets” installed on other servers in the organisation.

OEM is actually made up of two parts, and a database repository. The Oracle Management Service runs within a WebLogic domain and comprises of a Web Console (what we’d know as Enterprise Manager) and “Platform Background Services”, a set of background services that communicate with the target hosts and store the relevant information. The other part of OEM is the “Oracle Management Agent”, a server process thats installed on each monitored host that collects metrics to pass back to OMS and PBS, and executes tasks such as stopping and starting the target on behalf of OMS. OEM12cR2 Cloud Controls stores its metadata and monitoring data in a separate repository database, which can either be on the same server as OMS or on a separate machine – note that if you use a database instance that’s previously had Database Control enabled on it (as most of them have), you need to disable and remove it before you can use it for OEM’s own repository.

One of the main benefits of OEM12cR2 compared to standalone management consoles is that it manages the majority of Oracle’s server products – WebLogic Server, Oracle Database, Exadata, Exalogic, E-Business Suite and so on, though you need to read the small print as management covers more features in some products than others – we’ll get back to this point later on. At its best though, OEM12cR2 becomes your central monitoring point for all products (including some third party ones, via plugins), allowing you to monitor, manage, patch and maintain all of your servers from the one place.

As well as managing all hosts in one place, headline benefits of OEM12cR2 over “free” Fusion Middleware Control include:

• Monitor all BI Domains in one place, so you can see their versions, host types, patch levels etc
• Perform WebLogic lifecycle-type tasks such as patching the installation, packing and unpacking managed servers to move them between hosts, deploying test-to-production
• Define quality of service checks, create alerts for slow response times, hosts down etc
• Persist and store metrics, rather than only display them whilst you have the Metric screen open in your browser

Like the Oracle database though, Enterprise Manager comes with a number of extra-cost packs, including:

• Database Lifecycle Management Pack for Oracle Database
• Data Masking Pack
• Test Data Management Pack
• WebLogic Server Management Pack Enterprise Edition
• SOA Management Pack

and, of course, the BI Management Pack. So what do you get in the base version of OEM before you need to start paying for these packs? For all of the database, middleware and other targets, you can deploy agents, set alerts and define metric thresholds, and for Oracle Database specifically you can use the data movement features, view errors, use Advisor Central and so on, whereas the stuff you really want such as performance pages, wait event breakdowns and so on are extra cost. Same goes for WebLogic, with a small base-level set of functionality that’s pretty-much limited to discovering the WebLogic installation, then stopping and starting it, in other words what you get for “free” with Fusion Middleware Control. For BI, again you can display what you would normally see in Fusion Middleware Control (database and middleware licensed customers can use base-level Oracle Enterprise Manager at no extra license cost, so this would follow), but if you’re after anything else such as persisted metrics, service tests and so forth, figure on buying a few of the add-on management packs.

My article on OEM Grid Control 10gR4′s BI Management pack described the features that are still the core of OEM12cR2′s BI Management Pack, which at the time included the features below, as shown in the screenshot below.

• The ability to collect and record BI Server, BI Presentation Server, BI Cluster Controller and other BI target metrics, and define thresholds and events against those metrics
• The ability to connect to the BI repository database tables, to read for example the BI scheduler information about failed iBot executions and use it to alert you
• The ability to connect to the DAC repository, and then graph out ETL run information such as execution time, number of errors and so forth
• Record configuration settings, and then report on what’s changed for a target configuration compared to the previous settings

So now that the BI Management Pack is back with OEM12cR2, what do you get with it? Well you get everything that you had before, plus some new features:

• The ability to discover, and then monitor, Essbase installations
• All the new functionality around WebLogic (albeit with the requirement to license the WebLogic Management Pack)
• Compatibility with OBIEE 11g, along with continuing support for 10g

The screenshots below show some of these features in use, with the new EM12cR2 “Fusion” look and feel for the web console.

So how do we get EM12cR2 connecting to OBIEE, and make use of some of the new BI Management Pack features; also, how do we register an OBIEE installation with it, and how does it work with a BI Apps installation, or even about Exalytics? Come back tomorrow when we’ll cover off the installation and configuration parts of the product.