We have been seeing Excel Services/PowerPivot data refresh issues using SharePoint 2013 and Office Web Apps (OWA). 

Some customers have been running into issues when attempting to refresh data in a PowerPivot workbook on SharePoint 2013 using OWA server, and are seeing errors similar to the following:

"PivotTable Operation Failed
An error occurred while working on the Data Model in the workbook."

 or

 "External Data Refresh Failed
An error occurred while working on the Data Model in the workbook. Please try again.
We were unable to refresh one or more data connections in this workbook.  The following connections failed to refresh:
<Connection>"

First, a little background information on SharePoint 2013 and Office Web Apps:

When SharePoint 2013 is configured to use Office Web Apps (OWA), by default, Excel files stored on SharePoint are viewed in WOPI frames and can be edited via OWA.  However, when the SharePoint farm has been configured to use Excel Web Apps, the features available in Excel Services and Power Pivot will depend on how the Excel Web App server has been configured. 

Excel Web App runs in one of two modes: 

•  SharePoint view mode - Excel services is used to view workbooks in the browser.  For example, you will see a URL similar to the following: http://sharepoint/_layouts/15/xlviewer.aspx?id=http%3a%2f%2fsharepoint%2fPP%2fLCA%2520BI%2520-%2520Financial%2520Report%2520Usage.xlsx&Range='Financial+Report+Usage+Tracking'!A1&DefaultItemOpen=1

            We can see that the xlviewer.aspx is invoked to view the workbook. 

•  Office Web Apps Server view mode: Excel Web Apps is used to view workbooks in the browser.  For example, you will see a URL similar to the following: http://sharepoint/_layouts/15/WopiFrame.aspx?sourcedoc=%2FPP%2FLCA%20BI%20%2D%20Financial%20Report%20Usage%2Exlsx&action=view

            We can see that the WOPIFrame.aspx is invoked to view the workbook.  We can also see that Web Apps is rendering the workbook at the top of the browser window.  (See the screenshots above.)

Please see the following for an in-depth overview of the BI features in Excel Services available by each mode:

Overview of Excel Services in SharePoint Server 2013
http://technet.microsoft.com/en-us/library/ee424405.aspx 

When OWA Server view mode is used to view workbooks, the following BI features will not be available.

• Excel Web Access Web Part
• Refresh OData connections
• View and interact with Power View reports
• View and interact with PowerPivot data models
• Refresh PowerPivot data models
• Refresh data by using the Excel Services unattended service account
• Refresh data by using Effective User Name connections
• Kerberos delegation

We can use filters and slicers for PowerPivot workbooks if we suppress OWA from handling the .xlsx file type and force SharePoint to use SharePoint view mode.  We can do that by running the following command via PowerShell on the SharePoint farm:

New-SPWOPISuppressionSetting -extension xlsx -action view

New-SPWOPISuppressionSetting
http://technet.microsoft.com/en-us/library/jj219443

Once the suppression setting is applied to the farm, we can now work with slicers and refresh data.  We can still also edit the document in the browser with OWA!

Additional Resources:

Deploy Office Web Apps Server:
http://technet.microsoft.com/en-us/library/jj219455.aspx

Configure SharePoint 2013 to use Office Web Apps
http://technet.microsoft.com/en-us/library/ff431687.aspx
 
Overview of Office Web Apps and how they work on-premises with harePoint 2013
http://technet.microsoft.com/en-us/library/ff431685.aspx

I find myself explaining PowerPivot to customers over and over.  Most understand parts of how it works but not the whole story.  I will touch on two parts; PowerPivot for Excel and PowerPivot for SharePoint.

Part 1: PowerPivot for Excel

PowerPivot for SharePoint is nothing without the Excel Rich Client and the PowerPivot Add-in.  Pre-Excel 2013, the PowerPivot Add-in was a separate COM Add-in that you had to download.  In Excel 2013, the PowerPivot Add-in ships with the product.

Before you can do anything in SharePoint, you need to create a PowerPivot workbook in Excel.  To do this, open the PowerPivot tab choose Manage > From Database (choose your datasource, servername and database name).

You (potentially) will be pulling in a large amount of data into Excel.  I have seen people with Excel (PowerPivot) workbooks 1GB in size (which is huge since .xlsx (Open Office XML) files are zipped).  This large amount of data is essentially a database and you can see this in the workbook.  To view the Data Model, rename workbook file extension (.xlsx to .zip), open the workbook > xl > model > item.data. 

After you have pulled this data (essentially a database) into the PowerPivot environment, Excel can point to this database and build a PivotTable off it.  In the PowerPivot environment choose PivotTable > PivotTable

Excel is now pointing at the database (Data Model) embedded inside itself.  You can see for yourself via Excel > Data Connections > Properties.  In SharePoint (when we get there), we will call this the "Refresh in Browser".

Keep in mind there is still the PowerPivot Connection that can also refresh.  In SharePoint, we will call this the "Scheduled Data Refresh".  You can view this connection in Excel > PowerPivot > Manage > Existing Connections (choose the PowerPivot Data Connection) > Edit > Advanced.

There is still one more refresh (in SharePoint 2013 & SQL 2012 SP1) called the "Interactive Data Refresh".  I will touch on that later in the "What is PowerPivot for SharePoint?  Part 2".

At this point, we should have a beautiful Excel (PowerPivot) workbook that has a Pivot Table and Slicers and looks something like the below sample.

Now that we have a beautiful Excel (PowerPivot) workbook that functions wonderfully in the Excel Rich Client, we are ready to publish this to SharePoint!

Please see "What is PowerPivot for SharePoint? Part 2".

Now that we have a PowerPivot workbook in the Excel Rich Client, we are ready to publish that workbook to SharePoint.

In "What is Power Pivot for SharePoint? Part 1" I discussed the Data Model and how it is a database inside of Excel.  When you publish/upload a regular Excel workbook to SharePoint, that file gets chunked up into blobs and stored in the Content Database.  With a PowerPivot workbook the Excel workbook also gets stored in the Content Database, BUT the Data Model gets created and stored (after the initial (successful) refresh in browser) on the server running POWERPIVOT.

Note: In SharePoint 2010, the Analysis Service POWERPIVOT instance is going to be on a SharePoint server in the farm.  In SharePoint 2013, we recommend using a SQL Server off the farm.  Step by step install instructions here.

I am now going to touch on the three refreshes of PowerPivot 2013 and how they work.

1. Browser Refresh

When you open workbook and click a slicer, you are merely accessing the Data Model.  In other words, (in this example) the PowerPivot workbook stored in a SharePoint Report Library is going our to the SQL Server running PowerPivot > Databases > PowerPivot20Test20Workbook and is pulling whatever data (fresh or stale) that is currently in that cube.

If you login to SQL Management Studio > Analysis Services > Servername\POWERPIVOT > Databases you will see the Data Model (in this case PowerPivot20Test20Workbook_ae1bfdf63a44395b27a908c0ac0855_1f7abf568bb14cd4bb070b5eb858911c_SSPM):

Troubleshooting:  I have seen the refresh in browser fail periodically.  This was caused by not having the correct ASOLEDB and ADOMD.Net drivers (located here: for SQL 2008 R2 SP1 & SQL 2012 SP1) on the SharePoint Servers running Excel Services and/or the Claims to Windows Token Service not running on these machines/lacking proper permissions.

2. Scheduled Data Refresh

In SharePoint 2010, running the Scheduled Data Refresh is the only way to get fresh data into the PowerPivot workbook (this is designed to run once daily during "after business hours").  A timer job runs, the Data Model is updated with fresh data and workbook is republished back to SharePoint.  When you open the workbook and click on a slicer, the Browser Refresh occurs a you will see data from the time the Scheduled Data Refresh ran (data from the time the Scheduled Data Refresh ran that night).  In 2013, you are not 100% dependent on the "Scheduled Data Refresh" (if your backend data source accepts Windows Credentials; SQL or Analysis Services) to see fresh data.  In 2013, you can use the "Interactive Data Refresh" which I will touch on next.  If you want to refresh from a data source that does not accept Windows Credentials (Oracle, IBM (in some cases) SQL, etc.)  you will need to use the Scheduled Data Refresh in both versions of SharePoint (2010 & 2013).

For this functionality to be present is SharePoint 2013, you need to install the PowerPivot.msi on a SharePoint server.

Troubleshooting: For detailed instructions on how to set this up please follow this article. For help pulling from a data source that does not accept Windows Credentials, please see our blog.

 3. Interactive Data Refresh

The "Interactive Data Refresh" is only available in SharePoint 2013.  The "Interactive Data Refresh" passes the SharePoint user's credentials all the way to the backend and pulls real-time data into a PowerPivot workbook.  To do this, you need to choose Data > Refresh All Connections. 

You can pull data from a data source that accepts Windows Credentials (SQL & SSAS).  For steps on how to set this up, please see our blog.

 You may receive the following warning dialog box when opening a workbook using Excel Services in SharePoint:

Be careful. This workbook contains one or more queries that might be unsafe. Do you want to enable these queries?

To suppress this and other warnings from Excel Services, go to SharePoint Central Administration > Application Management > Service Applications > Manage service applications > Excel Services application > Trusted File Locations and click on the Address that contains the workbook.

In the External Data section, under Warn on Refresh, uncheck Refresh warning enabled and click OK.

Users who access the workbook will no longer be presented with a warning dialog box.

Issue:

The PowerPivot addin for Excel is used to create a PowerPivot model using SQL Server as a data source and the user manually uses the Excel connections screen to have the password be saved.

Later when the PowerPivot addin is used to change the model, like by changing some query text the model will fail to refresh.

Cause:

The model command text gets out of sync with what Excel has for the command text.  This is a known issue and may not be resolved anytime soon.

Workaround:

Using SQL Server authentication is not recommended for many reasons, Windows authentication to the SQL database is the recommended authentication mode.

Storing a login ID and password to a database is exceptionally insecure and a not recommended.  As the workbook gets passed around the password is visible to anyone who opens the workbook.

If you must use SQL authentication, it can be done in a much more secure way if you have SharePoint and PowerPivot for SharePoint.

1. Create the workbook with Excel same as before, but never check to "Save Password", you will be prompted whenever the password is needed.
2. Upload the file to a SharePoint PowerPivot Gallery document library  
3. In the Gallery click the "Manage Data Refresh" button
4. Configure a schedule to refresh the data, you can put the SQL userid and password in the schedule, these values are never readable by users
5. Have users use the workbook

The above method is vastly superior to saving the password in the workbook.  At no time is the password visible to a anyone, only the workbook author needs to know the password.  The data is only refreshed once per day or whatever interval is set in the schedule.  This is a large performance gain.  Without scheduled data refresh, the data refresh processes might be done by every user that opens the workbook multiple times, with scheduled data refresh it only happens once per time period.

Every once in awhile, you will get an error where you cannot Slice to the Data Model. 

"PivotTable Operation Failed
An error occurred while working on the
Data Model in the workbook.  Please try again."

If this is the case, you will want to make sure PowerPivot is Configured correctly.  To do this, follow the below instructions.

Turn on ULS logging:

Tick Excel Services Application > Least critical event to report to the trace log > Medium

On the machine running Excel Services, connect to the ULS log via ULS Viewer > Open From > ULS

Filter on Message > Contains > check server configuration

Then Stop and Start > Excel Calculation Services. After 30 seconds (or so) you will see an entry:

Once you see that, you know the Check Server Configuration took place.

Click on that event and then clear the filter via Edit > Clear Filters

You can then see all 4 checks:

I will go through all the steps below:

1. Check Administrator Access (AS2012SP1\POWERPIVOT): Pass.

If this fails, it will look similar to:

"w3wp.exe (SP:0x4968)        0x2BE0        Excel Services Application        Data Model        27        Monitorable        –> Check Administrator Access (SQL\POWERPIVOT): Fail."

If this fails, you will need to make sure the account running the PowerPivot Application Pool has access to the Analysis Services machine running PowerPivot. 

Below is where the PowerPivot Application Pool account needs to be present.

2. Check Server Version (sql\POWERPIVOT): Pass (11.0.3393.0 >= 11.0.2800.0).

I don't have an example of this error, cause no one has ever been on a build lower than SQL 2012 SP1.  If this fails, you need to make sure the Analysis Services machine running PowerPivot is at least on SQL 2012 SP1.

3. Check Deployment Mode (sql\POWERPIVOT): Pass.

If this fails, it will look similar to:

"w3wp.exe (SP:0x4968)        0x2BE0        Excel Services Application        Data Model        27        Monitorable        –> Check Deployment Mode (SQL\POWERPIVOT): Fail (Expected: SharePoint, Actual: Multidimensional)."

This means that you did not install Analysis Services in SharePoint Mode.  The above error shows Multidimensional Mode. 

Below is the correct configuration:

4. Check Server Configuration (AS2012SP1\POWERPIVOT): Pass.

If this fails, it will look similar to:

"w3wp.exe (SP:0x4968)        0x2BE0        Excel Services Application        Data Model        27        Monitorable        Check Server Configuration (SQL\POWERPIVOT): Fail (Uninitialized, ConfigurationError, InsufficientPermission, WrongDeploymentMode)."

This essentially is a summary of the above 3 steps; as you can see the Permission to the Analysis Services machine running PowerPivot were insufficient (account running PowerPivot Application Pool needs to be an Administrator to the /PowerPivot Instance) and the Deployment Mode was not SharePoint Mode.

Hopefully this helps. 

Here are the steps I performed to Consume a SharePoint List in Excel Services (SharePoint On Premise):

I created a list at this location:

http://sp/_layouts/15/start.aspx#/Lists/TomTestList/AllItems.aspx

I then called the listdata.svc from via Excel > Data tab > From Other Sources  > From Odata Feed > http://sp/_vti_bin/listdata.svc

I selected my List.

I then created a simple Pivot Table.

I went to Data > Connections > Properties > Definition tab > Authentication Settings > Use a stored account > (enter a SSS ID you have created).

NOTE: the “Set Credentials” account for this SSS ID needs to have access to the SharePoint List.

I changed the List (I added CHEESE).

I then clicked Refresh All Connections. As you can see CHEESE appeared (this wasn’t instantaneous.  It did take a little while, so be patient).

IMPORTANT NOTE:

When you consume a SharePoint List in Excel Services via the ListData.svc, a Data Model is created.

and the Data Model is built on the PowerPivot Server:

If you do not have a PowerPivot Server, the refresh will fail.

Also, if you are using an Office Web Apps 2013 Server, you will need to suppress .xlsx file types from being handled by OWA Server:

New-SPWOPISuppressionSetting -extension xlsx -action view

To reverse this:

Remove-SPWOPISuppressionSetting -extension xlsx -action view

Related Blog:

PowerPivot for SharePoint – Browser Refresh Fails (Data Refresh not supported in Office Web Apps)

http://blogs.technet.com/b/excel_services__powerpivot_for_sharepoint_support_blog/archive/2013/01/31/powerpivot-for-sharepoint-browser-refresh-fails-data-refresh-not-supported-in-office-web-apps.aspx  

You may receive the following error when running the PowerPivot Configuration Tool on a SharePoint application server running PowerPivot for SharePoint:

The user is not a farm administrator. Please address the validation failures and try again.

The user is required to be both a farm administrator and a site collection administrator in Central Administration.The error may occur if the user is not assigned to one of these groups.

To add the user as a farm administrator, go to Central Administration > Security > Manage the farm administrators group. Click on New > Add Users, enter the user, and click Share:

To add the user as a site collection administrator in Central Administration, go to Central Administration, click on the gear icon on the top right, and select Site settings. Then go to Users and Permissions > Site collection administrators. Enter the user and click OK:

After ensuring the user is in these two groups, the user should be able to run the PowerPivot Configuration Tool successfully.

Updated 9/22/2015

After migrating from a SharePoint 2010 site using Classic Mode Authentication(default) to SharePoint 2013 site in claims mode (default again) using database attach, you will run into a situation where document authors are not converted to the new claims identities by the migration process. This is not limited to Excel documents, but is more relevant to us as PowerPivot uses these IDs while performing specific actions.

Symptoms and Overview:

In the scope of PowerPivot, this is an issue as our timer jobs expect that the user names will be in the correct format when we need to utilize them. If they are not, you can run into a very specific issue where the PowerPivot Management Dashboard Processing Timer Job fails with a "User cannot be found." error.

“Microsoft.SharePoint.SPException: User cannot be found” at Microsoft.AnalysisServices.SPAddin.UsageProcessingTimerJob.PrepareUsageData

Some or all of your PowerPivot files likely have their author property set in the old Windows style name like “MYDOMAIN\someuser”. If the author property was set to a claims format it would look like this: “i:0#.w|MYDOMAIN\someuser” and your timer job would run without issues.

Workaround:

The "easy" workaround for this issue is to delete and re-upload the PowerPivot workbooks affected. The author field cannot be fully examined from the SharePoint UI, so it can be difficult to track down all of these workbooks. Also, this workaround does not always modify the document properly.
Attached to this post (very bottom) is sample code that shows one way to list all of the workbooks for a PowerPivot service application that have an author property in the Windows style format (these files will throw and error when the "PowerPivot Management Dashboard Processing Timer Job" runs).

We have also run into scenarios where the timer job will fail, but our detection tool will not pick up the improperly formatted user name in the database. The workaround of re-uploading or modifying via script is the same regardless.

In the second scenario, in the ULS logs you will see the timer job iterate through all of your PowerPivot workbooks and fail on a specific book. This stops the timer job in its tracks and it will not proceed:

OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        General         g3ql        Verbose GetUriScheme(/site/subsite/anothersubsite/PowerPivotGallery/workbookname.xlsx)
OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        Security         ahluw        Verbose        Entering: GetByLoginNoThrow(domain\username)
OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        Security ahluz        Verbose        Exiting: GetByLoginNoThrow(domain\username)
OWSTIMER.EXE (ServerName)        0x879C        PowerPivot Service        Usage        99        High        EXCEPTION: Microsoft.SharePoint.SPException: User cannot be found.     at Microsoft.SharePoint.SPUserCollection.get_Item(String loginName)     at Microsoft.AnalysisServices.SPAddin.UsageProcessingTimerJob.UpdateDocumentIfNeeded(SPFile file, SqlDataReader reader, SqlCommand updateCommand)     at Microsoft.AnalysisServices.SPAddin.UsageProcessingTimerJob.UpdateDocuments(GeminiServiceApplication application, SqlConnection conn1, SqlConnection conn2)     at Microsoft.AnalysisServices.SPAddin.UsageProcessingTimerJob.PrepareUsageData(GeminiServiceApplication application)     at Microsoft.AnalysisServices.SPAddin.UsageProcessingTimerJob.Execute(Guid targetInstanceId)

In a working scenario, you would see the following for each workbook. Take note of the claims identifier prior to the domain:

OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        General g3ql        Verbose GetUriScheme(/site/AnotherWorkbook.xlsx)        
OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        Security         ahluw        Verbose        Entering: GetByLoginNoThrow(i:0#.w|domain\username)
OWSTIMER.EXE (ServerName)        0x879C        SharePoint Foundation        Security         ahluz        Verbose        Exiting: GetByLoginNoThrow(i:0#.w|domain\username)

After this entry, you will see normal traffic and the process moving on to the next document or ending the sequence if there are no other documents to parse.

** THIS SAMPLE CODE AND POWERSHELL SCRIPT ARE PROVIED AS IS, WITH NO WARRANTEES OR SUPPORT OF ANY KIND, USE AT YOUR OWN RISK **

PPTimerUserError.exe Usage:
Extract the zip and navigate to the following path:
DRIVE:\filelocation\PPTimerUserError\PPTimerUserError\bin\Debug
Rename PPTimerUserError.bak to PPTimerUserError.exe so you can run the program (.bak so zip scanners don't block the file)

The .exe must be run from one of the farm SharePoint 2013 servers as an administrator account that has access to your SharePoint Service Application databases. The output is tab delimited so you can open with Excel to sort and filter as needed.
Usage: PPTimerUserError <DB Server> <PowerPiviot DB Name>
Example: PPTimerUserError warrenr-ws1 DefaultPowerPivotServiceApplicationDB-57bc03ae-bd41-4102-ab6f-f03201e3b583 

Output will be placed here: C:\temp\PPTimerAuthorReport.txt  and PPTimerAuthorReportDetails.txt

You will find the information you need in the detailed report (PPTimerAuthorReportDetails.txt).

Find all of the documents that error with "User cannot be found".

You then have 2 options for fixing this. You can either download the document, delete it from the library and re-upload it, or we have built the following script to auto modify the document with an ID of your choosing:

**USE AT YOUR OWN RISK **

$web = get-spweb http://SITECOLLECTIONURLHERE/sites/MORE
$list = $web.lists['LIBRARY NAME']
$user = get-spuser -Web $web -limit all | ? {$_.userlogin -match "username"}
$sUser = $user.id.ToString() + ';#' + $user.DisplayName
$rItem = $list.items | ?{$_.name -eq "workbookname.xlsx"}
write-host "Current Author is:" $ritem.properties["vti_author"]
$ritem["Author"] = $sUser
$ritem.properties["vti_author"] = $user.loginname
$ritem["Editor"] = $sUser
$ritem.UpdateOverwriteVersion()
write-host "Updated Author is:" $ritem.properties["vti_author"]

You will need to modify the highlighted fields to fit your situation. Also note that the "username" field is literally that, just a user name. Do not input a user in "domain\username" format. You can run this script line by line as you test and validate its functionality. Lastly, when you run this, you may also see another PowerShell window open with some messages about snapshots. Just let it run. It is the GallerySnapshot process running manually due to the document update.

A note about the above script. We decided not to expand on this script as each person's usage may be different. Feel free to use this skeleton script in any way fits your situation (put it in a loop for a library or document type etc….) or even modify it to detect the current author and change it to the claims ID of the same user.

Please make sure you test this script in some kind of test farm thoroughly!!

PPTimerUserErrorv1.zip

When you create a data connection to an Analysis Services data source in Microsoft Excel, the Microsoft Analysis Services OLE DB Provider for Microsoft SQL Server (MSOLAP) is used for the data connection. Each version of Analysis Services has its own MSOLAP provider version. Here is a list of versions:

MSOLAP.4 – SQL Server 2008
MSOLAP.5 – SQL Server 2012
MSOLAP.6 – SQL Server 2014

Read the following article for more information about using the correct version of MSOLAP:

https://support.microsoft.com/en-us/kb/2735567

Excel will use the version of the MSOLAP provider installed on the client machine. In the following example, Excel has configured MSOLAP.5 as the data provider in the data connection string:

You may have multiple versions of the MSOLAP provider installed on the client machine. How does Excel know which MSOLAP version to use? It looks for the version that is configured in the registry. So even though you have installed MSOLAP.6 on your machine in order to connect to Analysis Services 2014, Excel may configure the connection to use MSOLAP.5 in the connection string instead. This will cause issues as you cannot use MSOLAP versions that are earlier than the data source version.

You can fix this by changing the MSOLAP version in the registry keys below, which define what Excel uses to connect to Analysis Services. The location of the registry depends on whether Microsoft Office is an MSI or Click-to-Run (C2R) installation, and whether it is 32-bit or 64-bit.

Office 32-bit MSI
 
HKEY_LOCAL_MACHINE\SOFTWARE\Classes\Wow6432Node\CLSID\{308FF259-8671-4df4-B66C-9851BFACF446}\ProgID\(Default)

Office 64-bit MSI

HKEY_LOCAL_MACHINE\SOFTWARE\Classes\CLSID\{308FF259-8671-4df4-B66C-9851BFACF446}\ProgID\(Default)

Office 32-bit C2R

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office\15.0\ClickToRun\REGISTRY\MACHINE\Software\Classes\Wow6432Node\CLSID\{308FF259-8671-4df4-B66C-9851BFACF446}\ProgID\(Default)

Office 64-bit C2R

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office\15.0\ClickToRun\REGISTRY\MACHINE\Software\Classes\CLSID\{308FF259-8671-4df4-B66C-9851BFACF446}\ProgID\(Default)

Here is an example of Office 32-bit C2R configured with MSOLAP.5:

To find out whether your installation is MSI or C2R, in Excel go to File > Account. If you see an Office Updates section, the installation is C2R:

If there is no Office Updates section, then it is an MSI installation:

 To find out if Excel is 32-bit or 64-bit, click on About Excel in the same Accounts screen and it will show you in the dialog box at the top:

When you open a workbook in the browser using Excel Services and get an error, or when a scheduled data refresh fails in PowerPivot for SharePoint, you may see the following error in the ULS logs:

Retrieving the COM class factory for component with CLSID {BDEADF26-C265-11D0-BCED-00A0C90AB50F} failed due to the following error: 800703fa

If you see this specific error, you can fix the error using the following steps:

1. Open Internet Information Services (IIS) Manager.

2. Expand SharePoint Web Services in the left menu.

3. Click on the first ID under SharePoint Web Services and then click on Content View at bottom of the center menu.

4. Select each ID until you see ExcelService.asmx (for Excel Services error) or SSASMidTierService.svc (for PowerPivot error) in the center menu.

5. Click on Advanced Settings on the right menu, note the value in Application Pool, and click Cancel.

6. Select Application Pool in the left menu and select the application pool you noted in the previous step.

7. Click on Advanced Settings on the right menu.

8. Look for the setting Load User Profile, set the value to True, and click OK.

9. Open a Command Prompt and run iisreset to restart IIS.

Overview:

This blog post contains a tool that will query all content databases on a SQL server looking for files of a given extension and size, it will output a report file that can be loaded into Excel for analysis.

PowerPivot can lead to some large workbooks being stored on your SharePoint farms, the tool in this blog post can be used to scan multiple content databases looking for files based on size and file extension.

Disclaimer:

Use this tool at your own risk, there is no support, guarantees or warranties for this sample program.  This sample program will query all databases on you SQL server looking for content databases and when found doing a read only query against them.

How to use:

Your Windows credentials must have read access to the master database on the SQL server and read access to all content databases on the server.

You must have a c:\Temp directory on the computer (for the output file)

A complied exe can be found in the SPFileSizeQry\SPFileSizeQry\bin\Debug folder

The output text file has these columns [DBName] [DirName] [LeafName] [Size] [TimeLastModified]

Usage:

SPFileSizeQry.exe <SQLServer> <Extension to query> <min size in bytes to report>
example: SPFileSizeQry MySqlServerAddr xlsx 50000000
output at C:\temp\SPFileSizeReport.txt 

The source and compiled program are in the zip file attached to this blog post.

SPFileSizeQry.zip

We have been seeing some issues with customers using PowerPivot workbooks as a source file in new Excel PowerPivot files.  There are a variety of errors that can arise from this process, and this blog is intended to provide you with some troubleshooting steps to resolve the issue.

You may or may not be aware, but it is possible to use a PowerPivot workbook as a data source for other data applications.  For instance, one can use Excel and connect to a PowerPivot file on a SharePoint site as a PowerPivot data source.  There may be a file on SharePoint site named BIFinance.xlsx.  When creating a new PowerPivot workbook, we could specify our data source as the following:

http://sharepointsite/PowerPivot Gallery/BIFinance.xlsx

Fig 1: Using PowerPivot workbook as a data source

In a nutshell, when the data source is the PowerPivot Mid-Tier hosted PowerPivot workbook, the connection goes through a Redirector service on the SharePoint server and is ultimately routed to the SQL Server Analysis Services PowerPivot instance.

For more information on how to use PowerPivot workbooks as a Data Source, please see the following video:

Using PowerPivot Workbooks as a Data Source
http://technet.microsoft.com/en-us/sqlserver/dn151361.aspx

A few sample errors that may arise when attempting to connect to the PowerPivot workbook as a data source:

• Failed to connect to the server. Reason: No error message available, result code: DB_SEC_E_AUTH_FAILED(0x80040E4D)
• Unable to connect to data source. Reason: Access denied. You either made a mistake typing in your User ID and/or Password, or you do not have permission to access the database server.
• The following system error occurred: Invalid class string
• PowerPivot Web service returned an error

If you see the "Access Denied"or "Invalid class string" errors, try the following:

1. The SharePoint Web application is setup with Kerberos.  If you are seeing the DB_SEC_E_AUTH_FAILED(0x80040E4D) error, then the web application is more than likely set up with Kerberos. 

a. Please follow the steps in the following article to resolve the issue.  You will need to modify the web.config file on the web front end servers.
  http://blogs.msdn.com/b/johndesch/archive/2012/04/23/using-powerpivot-workbooks-from-a-mid-tier-server-configured-for-kerberos-authentication.aspx

2. The client machine making the call or the SharePoint servers do not have updated Microsoft SQL Server Analysis Server OLEDB and ADOMD.Net drivers. You may need to check with your SharePoint administrator to determine what version of PowerPivot is installed in your SharePoint environment.

a. If using SQL Server 2008 R2 PowerPivot:
http://www.microsoft.com/en-us/download/details.aspx?id=30440

b. If using SQL Server 2012 PowerPivot:
http://www.microsoft.com/en-us/download/details.aspx?id=35580

3. Make sure the SharePoint Web application where the workbook is stored does not have multiple bindings in IIS. 

a. Open IIS on the SharePoint Web Front End servers and select the SharePoint web application and click on Bindings. 

b. If there is more than one this will fail.

c. If your web application needs more than one binding, you may extend the web application in SharePoint. Please see the following article on how to extend the web application in SharePoint.

Extend a Web application (SharePoint Server 2010) – http://technet.microsoft.com/en-us/library/cc261698(v=office.14).aspx

4. The user that is making the connection needs to have read access the root of the web application. Go to Central Administration > Application Management > Manage web applications > User Policy and give read access to user.

5. Restart Excel Services on all the application servers running Excel Services.

6. If you see this error: "XML parsing failed at line 1, column 1: Incorrect document syntax." Please review the following article:

http://blogs.technet.com/b/excel_services__powerpivot_for_sharepoint_support_blog/archive/2015/01/28/error-when-using-powerpivot-workbook-as-a-data-source-quot-xml-parsing-failed-at-line-1-column-1-incorrect-document-syntax-quot.aspx

7. If you see this error: "PowerPivot Web service returned an error", try the following:

a. The PowerPivot workbook being used as a data source needs to be in Trusted File Locations in the Excel Services settings. Go to Central Administration > Application Management > Manage service applications. Select the Excel Services application, click on Trusted File Locations, and add the location of the workbook.

b. The PowerPivot System Service account (PowerPivot application pool account) needs to be an administrator of the PowerPivot Analysis Services instance. Connect to the PowerPivot instance in SQL Server Management Studio, right-click on the instance, select Properties, select Security and add the account.

The steps above should resolve most of the errors that are seen when attempting to use a PowerPivot workbook as a data source within Excel PowerPivot.

You may receive the following error when attempting a data refresh using PowerPivot for Excel:

Memory error: Allocation failure : Not enough storage is available to process this command. . If using a 32-bit version of the product, consider upgrading to the 64-bit version or increasing the amount of memory available on the machine.

The current operation was cancelled because another operation in the transaction failed.

This error may occur when using the 32-bit version of Excel, which enables you to work with up to 2GB of data in memory when using PowerPivot for Excel. The PowerPivot add-in runs as an extension of Excel and the PowerPivot Vertipaq in-memory engine loads within the same process space. As 32-bit Excel is limited to a 2GB virtual address space, once you start adding up all of the uses of that virtual address space, including Excel, all of the add-ins, and the in-memory database itself, the largest PowerPivot workbook that you can create on a 32-bit machine may actually be much lower than 2GB.

Installing the 64-bit versions of Excel 2010 and PowerPivot for Excel 2010 would increase the 2GB limit to 4GB. Furthermore, if you use 64-bit Excel 2013, there is no imposed file size limit, so you would be able to fully utilize the amount of memory you have installed when using 64-bit PowerPivot for Excel 2013. Alternatively, you could try reducing the imported datasets and tables by applying filters in the Table Import wizard to avoid hitting the limit when using 32-bit Excel and PowerPivot.

You may experience Excel Services crashing continuously after applying a patch to SharePoint 2013.

The issue can be resolved by following these steps:

1. Make sure that the account running Excel Services is part of the groups WSS_WPG and WSS_ADMIN_WPG on the SharePoint servers running Excel Services.

2. Open the SharePoint Management Shell and run the following command:

psconfig -cmd secureresources

3. Run iisreset in the command prompt. This will reset the registry permissions so that the WSS_WPG and WSS_ADMIN_WPG groups have read access to the following registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office Server\15.0

Excel Services should now run successfully without crashing.

Recently we have seen multiple instances where custom 3rd party tools/scripts are being used to bypass the standard daily scheduled data refresh limit for PowerPivot workbooks. The main reason we have seen these instances is because these tools have been causing performance issues for SharePoint Web Front End servers that were not accounted for in their design and/or implementation into the subject farms. I first want to note that we do not support these tools in any way and highly recommend against using them at all. If you need to refresh your PowerPivot workbooks more than once a day, then you likely need to find a different BI solution to display your data. Excel Services and PowerPivot for SharePoint are not meant to be a “live feed” of data.

In these cases, we found that the tools and scripts update the workbooks from another server/machine outside of the farm and re-upload them to SharePoint. As in most cases where PowerPivot is used, the library where the documents were stored was a PowerPivot Gallery as it provides the best user experience visually and functionally. Due to the fact that the documents were stored in a gallery, the massive front end load was caused from the PowerPivot gallery snapshots being triggered on the workbooks. Every time a user (program, script or otherwise) updates or uploads a workbook to a gallery, this snapshot process is triggered. The process engages the workbook after upload and literally creates images for every supported sheet in the workbook and uploads them to the metadata of the file.

Now, if we take this as an example: If I have a PowerPivot gallery with fifty 30MB workbooks in it and I use my script to refresh all of them and then re-upload them to the site, not only did I just slam a single WFE with 1.5 GB of uploads, but I have potentially caused 50 getsnapshot processes to launch at the same time and start processing. Each of those processes reaches out to the site at the same time to create an image file and then upload it to the document. You can also see how this can affect more than one WFE at the same time.

The other difficult part about this is that a solution like this can be implemented by any user that has access to the farm. Administrators may not even know it exists until it is too late. Most of these update solutions can be run and triggered from a CLIENT machine. They do not require servers or code level access to function. They generally access the site via the same interface that other software (like Office for instance) would use.

Lastly, if the PowerPivot gallery is configured with versioning, this can cause some really bad issues with version limits. You may find these documents becoming quickly unusable or causing corruption in your content database if too many versions are created. They also cause massive storage issues on the database side when they get out of hand.

So, some things to think about:

1. Do not use custom solutions to update your PowerPivot workbooks more than once per day. (It is not supported)
2. If you do, do not store the workbooks in a PowerPivot Gallery.
3. Do not configure versioning on the library where the workbooks are stored (or limit the versioning severely).
4. If you need to refresh BI data more frequently, re-think your BI data strategy.

What can you do if you need data more frequently?

1. You could “upsize” your data models. Effectively this involves moving a data models created in PowerPivot to Analysis Services. You can then use SQL jobs to update the data as frequently as you wish. After that is configured, you could create Excel workbooks to hit the new model and “Refresh on Open”. This would effectively give you “fresh” data every time you open the workbook. Again, it would not be a live stream, but it would give you a supported method to retrieve data on a more frequent basis.
   

   Upsizing PowerPivot 2013 Workbooks to SSAS for Knowledge Workers

   • Note: This workaround may not make sense if you have very small workbooks.
2. If the data source is SQL Analysis Services, you could use flat out use Excel rather than PowerPivot and again configure the workbook to “Refresh on Open”. This solution will likely cause a little more load on the application servers as well as the Analysis Services data sources, but it is fully supported and a much more reliable method to retrieve data “on the fly”.
3. Consider other products that integrate with SharePoint such as SQL Reporting Services or PerformancePoint to display the data.

Just to define what we mean when we say something “Is not supported”. We at Microsoft will not provide support to fix custom solutions and/or scenarios that have been deemed “out of bounds” by our product teams. This does not mean that these solutions will not work, but if you run into a problem due to one of these solutions, we will not be able to assist beyond a best effort to get the farm back into a working state after the customizations have been removed and/or disabled.

If you question whether a solution you want to implement is supported, please contact support and we can take a look at let you know.

Scenario:

You are installing PowerPivot for SharePoint 2016 (or 2013) using SQL 2016 and you notice that you are no-longer restricted to having the instance name “POWERPIVOT” to make it function. As such, you decide to install PowerPivot Analysis Services on the default instance for the sake of simplicity. You continue with your Office Online Server and SharePoint configuration and proceed with testing your first PowerPivot workbook. The workbook opens without issues, but as soon as you attempt to interact with the PowerPivot workbook, you receive the following error: “Cannot locate a server to load the workbook data model”.

Problem:

At the moment, there is a known issue causing issues with SSPM streaming between OOS/Excel Services and PowerPivot when it is installed on the default instance. There is no known fix date for this issue.

You will notice a few symptoms that are user facing and on the back end. First, as stated above you will be unable to refresh (browser or scheduled) or interact with Excel workbooks that contain PowerPivot data models. On the back end, you will notice that no data models are being created inside of the PowerPivot Analysis Services instance, but all connectivity to the instance will seem without flaw as you will have no issues connecting with SQL Management Studio or PowerShell.

Workaround:

For now, you will be required to install PowerPivot Analysis Services on a named instance of SQL. The name of the instance is not relevant as long as it is within the constraints of the current naming limitations for all SQL named instances. Installing on a named instance causes no security issues with test or production deployments. When this issue is resolved, if you wish to migrate your PowerPivot Analysis Services installation to the default instance, you can do a side-by-side installation change your configuration at that time.

Update: Official KB Published: https://support.microsoft.com/en-us/kb/3199726

Scenario:

You configure Office Online Server with SharePoint 2013 or SharePoint 2016 and have configured Kerberos Constrained Delegation to a data warehouse to pull business intelligence data for your users. You have two Active Directory domains in separate forests in a 2-way transitive trust configuration (We will call them DomainA in Forest 1 and DomainB in Forest 2). The SharePoint farm, data source and Office Online Server sit in Domain A, while only a specific set of users reside in Domain B. Users have no issues logging into the site and interacting with SharePoint.

Users that sit in DomainA configure Excel workbooks and are able to load them into SharePoint, render them in the browser and refresh without any issues. This proves KCD is working. Users from DomainB come over to the SharePoint site and attempt to refresh the same workbooks and receive the following failure:
“The data connection uses Windows Authentication and user credentials could not be delegated. The following connections failed to refresh: XXXXXXXX”

The first assumption is that there is an issue communicating between domains because we have proven that the KCD is functional. You may have also found that this worked (or still works) with Excel Services in SharePoint 2013. You will find (through multiple means) that you can query users from the trusted domain without issues. All diagnosis and troubleshooting efforts will show that communication seems to be correct.

Digging Deeper:

If you dig deeper into this issue and pull Office Online Server logs, you will see the following chain in this scenario:

11/04/2016 11:25:38.51 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services 6k5r Verbose CredentialsProvider.GetCredentials: Getting credentials for external source: EXAMPLEDATASOURCE, with CredentialsMethod Integrated
11/04/2016 11:25:38.51 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a2pb1 Verbose CredentialsProvider.TranslateSid: Trying to translate SID: s-1-0-00-000000000-0000000000-0000000000-0000
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a2pb2 Verbose CredentialsProvider.TranslateSid: Searching domain: DomainA.lab
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a2pb2 Verbose CredentialsProvider.TranslateSid: Searching domain: DomainA.lab
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a2pb4 Medium CredentialsProvider.TranslateSid: SID not found on any domain

The above causes this:
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a1tj8 Medium CredentialsProvider.TranslateSidToUpn translation elapsed 285ms
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services a1tj0 Verbose CredentialsDelegation.UpnLogon: upn is null (ok on DataCenter)

Thus, causing the Windows Identity Translation to fail:
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services c9la Medium CredentialsProvider.GetCredentials: Failed to get WindowsIdentity.
11/04/2016 11:25:38.79 w3wp.exe (0x4D94) 0x193C Services Infrastructure Services Infrastructure Logging ai2vt Verbose An event was fired: id: NoIntegratedConnectionsAllowed, eventId: 5252, eventType: Warning, eventMessage: Credential delegation failed because Excel Services Application was unable to obtain a Windows Identity.  [Session: <<TRUNCATED>> User: i:0#.w|domainb\user1]
11/04/2016 11:25:38.81 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services oyyh Verbose CredentialsProvider.ProcessCancel: Processing cancel of the credentials for external source with data connection name: EXAMPLEDATASOURCE
11/04/2016 11:25:38.81 w3wp.exe (0x4D94) 0x193C Excel Online Excel Calculation Services ajhkx Medium SessionUser.EnsureCredentialsForExternalData: Failed to get credentials. User=i:0#.w|domainb\user1, ConnectionIndex=0

You will also notice that there are no calls to the other forest from the Office Online server in network traces (at least in relation to this stack).

Diagnosis and Workarounds:

To perform any data refresh action, Office Online Server requires the Windows Identity of the user which is gathered by the Claims to Windows Token Service. You will notice in the logs above that Office Online Server did not even try to query DomainB. The product team confirmed that this is a product limitation. At the moment, Office Online Server will only attempt to query domains located within the same forest as itself.

From the domain level, you can workaround this issue by moving both domains into the same forest. From the Excel Online level, you can refresh data using the Secure Store instead of the Authenticated Users account. This will effectively bypass the need for Office Online Server to retrieve user credentials from the other domain.

NOTE: This post is specific to Office Online Server on-premise and does not reflect on or speak to the capabilities of Office Online, Office 365 or SharePoint Online.

The PowerPivot Gallery has had some functionality and supportability changes between SharePoint 2013/2010 and SharePoint 2016. This is due to the fact that Excel Services functionality has moved to Office Online Server. While the basic premise of the gallery remains the same and some of the same rules apply, there are some changes that you should be aware of that will require your attention while troubleshooting or configuring a new farm. This article is purely informative and is not meant to include any troubleshooting steps. We will publish more information on a case-by-case basis for relevant issues.

What is new?

1. Gallery Snapshots no-longer support PowerView in any form (this includes Power View reports imbedded in Excel/PowerPivot workbooks AND .RDLX files built from PowerPivot data models stored in the gallery). Here is an example of what you will see if you load a workbook with Power View sheets included:
   
   • The event receiver that detects the file upload has been modified to accommodate for this. When it detects an XLSX file, it flags “false” and when it detects anything else, it flags true per the below ULS entry examples:
     
     • We recommend loading your RDLX files into a regular document library to avoid any confusion with users.
     • Also, we recommend publishing PowerPivot workbooks with PowerView sheets to PowerPivot galleries by sheet to avoid snapshots being taken of the erroneous sheets (see the following blog for this “workaround”).
       • https://blogs.technet.microsoft.com/excel_services__powerpivot_for_sharepoint_support_blog/2014/08/14/power-view-sheet-causes-thumbnail-error-in-powerpivot-gallery/
2. You will now see charts and sheets have separate snapshots in the snapshot ribbon, for instance a sheet with multiple charts will be shown, then each chart will have its own snapshot preview.
   
3. Gallery Snapshots are now handled by Office Online Server. The old gallerysnapshot.exe and getsnapshot.exe processes that used to take the snapshots no longer exist. Office Online Server utilizes the Excel REST API to render and load the snapshots to the documents when they are uploaded to the library or modified. Also, it is done in the context of the user that performs the action rather than a service account.
   • For administrators: You will find log entries for these actions on the SharePoint Web Front End and Office Online Server ULS logs during the upload or update of the PowerPivot file in question.
     • If a VALID .xlsx document is uploaded, you will see an entry similar to the following ULS entry as the first call to the “ExcelRest.aspx” as the users login:
       • Name=Request (GET:http://<SITEURL>/_vti_bin/ExcelRest.aspx/<LIBRARY>/<DOCUMENT>.xlsx/model/Sheets
4. Because Gallery Snapshots are now taken in the context of the user that performs the update action to the document, it is a REQUIREMENT that the SharePoint User Profile Service and User Profile Sync is running and fully functional. It also must be associated with the web application of which the PowerPivot Gallery is hosted in.
5. It also requires that the Claims to Windows Token Services be running on your SharePoint Web Front End servers, but if you are in MinRole configuration, this should be running anyway.

What is the same?

1. The gallery continues to only have full support for PowerPivot documents.
2. The same classic views are still included with the gallery.
3. The PowerPivot Gallery (and all supporting features) are still only supported with Windows Authentication (SAML and Anonymous are not supported under ANY circumstances).
4. You can still add BISM content types and create Power View reports from PowerPivot data models and BISM connections directly from the PowerPivot Gallery GUI.
5. The Scheduled Data Refresh interface remains the same and you access and interact with it the same way as in SharePoint 2010 and 2013 via the PowerPivot Gallery.

This article is specific to PowerPivot in SharePoint 2016 On-Premise with Office Online Server. If you are looking for information on Gallery Snapshots for SharePoint 2013 or SharePoint 2010 with PowerPivot, please see our other blog posts on this subject:

• https://blogs.technet.microsoft.com/excel_services__powerpivot_for_sharepoint_support_blog/2012/12/27/powerpivot-for-sharepoint-gallery-thumbnails-not-rendering/
• https://blogs.technet.microsoft.com/excel_services__powerpivot_for_sharepoint_support_blog/2012/12/06/troubleshooting-the-powerpivot-gallery-thumbnails-not-rendering-invoking-and-analyzing-the-getsnapshot-exegallerysnapshot-exe/
• https://blogs.technet.microsoft.com/excel_services__powerpivot_for_sharepoint_support_blog/2015/03/25/with-powerpivot-2012-for-sharepoint-2013-thumbnails-may-not-always-generate-properly-in-the-powerpivot-gallery-for-non-powerpivot-workbooks-red-x/

If you have not already installed SQL 2012 SP3 CU6 or CU7 on your PowerPivot for SharePoint Analysis Services instance and/or the MSI DO NOT PROCEED WITH THE INSTALLATION!

A recent product issue was introduced with SQL 2012 SP3 CU6 and CU7 into PowerPivot for SharePoint 2013. When users attempt to schedule or modify a refresh schedule for a PowerPivot workbook, they will be presented with the following error:

You will see a stack similar to the following in the ULS logs when this happens:

00/00/2017 00:00:00.00 w3wp.exe (XXXXXXXXX) 0x91D4 SharePoint Foundation General 8nca Medium Application error when access /_layouts/15/PowerPivot/ManageDataRefresh.aspx, Error=The resource object with key ‘DataRefresh_Warning_DataSources’ was not found.   at <<TRUNCATED STACK>>

00/00/2017 00:00:00.00 w3wp.exe (XXXXXXXXX) 0x91D4 SharePoint Foundation General b9y9 High Failed to read resource file “<C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\15\Resources\PowerPivot_Pages.en-US.resx>” from feature id “(null)”.

00/00/2017 00:00:00.00 w3wp.exe (XXXXXXXXX) 0x91D4 SharePoint Foundation General b9y9 High Failed to read resource file “<C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\15\Resources\PowerPivot_Pages.resx>” from feature id “(null)”.

The product team is working on a fix for this issue and it will be released in a future cumulative update.

For now, the workaround is as follows (ONLY IF YOU HAVE ALREADY installed the patch!!!):

1. Navigate to the following file location on your SharePoint servers: “C:\Program Files\Common Files\microsoft shared\Web Server Extensions\15\TEMPLATE\LAYOUTS\PowerPivot”
2. Create a backup of “ManageDataRefresh.aspx”
3. Open “ManageDataRefresh.aspx”
4. Find the following:

   <tr>
      <td colspan="2" style="height:20px;vertical-align:middle;">
          <asp:Label style="margin:5px;" ID="Literal3" runat="server" Text="<%$Resources:PowerPivot_Pages, DataRefresh_Warning_DataSources %>" />
      </td>
    </tr>

5. Replace it with the following:

    <%--
       <tr>
           <td colspan="2" style="height:20px;vertical-align:middle;">
              <asp:Label style="margin:5px;" ID="Literal3" runat="server" Text="<%$Resources:PowerPivot_Pages, DataRefresh_Warning_DataSources %>" />
           </td>
       </tr>
    --%>

6. Save the file.
7. IISReset the server.
8. Test the results once every server has been modified.

This blog will be updated once the fix is released or another workaround is provided by the product team.

Note that this is specific to PowerPivot for SharePoint 2013 and SQL 2012 SP3 CU6 and CU7. This issue should not affect any other products and/or Analysis Services stand-alone installations. The issue appears to be limited to the PowerPivot add-in for SharePoint (sppowerpivot.msi) which happens to be included with the CU installer.