The documentation is online in this article, but it doesn't clearly show examples or provide enough context, so I will try to do that here.https://technet.microsoft.com/en-us/library/ms186839(v=SQL.105).aspx

1. The first RPC setting is mainly for a legacy feature called Remote Server, which is documented here. https://msdn.microsoft.com/en-us/library/ms190187.aspx

You probably will not be using remote servers in SQL Server 2005 -SQL Server 2014 versions.

If you do happen to use "remote servers", the RPC setting enables a certain security feature. If it is turned off, when the second server (the one receiving the login attempt from the first server to which the client connected) tries to validate the remote login, it fails with error

18482 “Could not connect to server '%.*ls' because '%.*ls' is not defined as a remote server. Verify that you have specified the correct server name. %.*ls.”

Kudos to my peer Ignacio Alonso Portillo for helping figure out this legacy option.

2. The RPC OUT  setting is very pertinent to linked servers on SQL Server 2005 -SQL Server 2014 versions.

Think about an RPC (Remote Procedure Call) as being a stored procedure being run remotely from Server 1 to linked Server 2 "myserver".

There are various syntaxes, the first using 4 part naming servername.databasename.schemaname.procedurename , the second here using EXECUTE(databasename.schemaname.procedurename ) AT servername convention.

We can test with a simple procedure sp_helpdb that lives in master database, and will list out the databases and some of their key properties.

• EXEC [myserver].master.dbo.sp_helpdb
• EXEC ('master.dbo.sp_helpdb') AT myserver

These kind of remote stored procedure calls will be blocked unless RPC OUT is set to True.

Msg 7411, Level 16, State 1, Line 1  Server 'myserver' is not configured for RPC.

You can set it to True or False in the linked server's properties (a right click menu in from the Linked Server Name in SQL Server Management Studio) .

Or you can set it by calling the stored procedure to set the server options like this test:

I hope this helps someone out there. Let me know if not. ~Jason

When we deploy the package to SSISDB in SQL Server 2012 RTM or SP1 (11.0.3153.0) and run it, it returns the following error, because the SSIS 2014 package format is not downgradable to SSIS 2012 version format.

 "The package failed to load due to error 0xC0010014

""One or more error occurred. There should be more specific errors preceding this one that explains the details of the errors. This message is used as a return value from functions that encounter errors."". This occurs when CPackage::LoadFromXML fails.

   at Microsoft.SqlServer.Dts.Runtime.Package.LoadFromXML(String packageXml, IDTSEvents events)

   at Microsoft.SqlServer.Dts.Runtime.Project.LoadPackage(IProjectStorage storage, Package package, String streamName, IDTSEvents events)

   at Microsoft.SqlServer.Dts.Runtime.PackageItem.Load(IDTSEvents events)

   at Microsoft.SqlServer.Dts.Runtime.PackageItem.LoadPackage(IDTSEvents events)

   at Microsoft.SqlServer.IntegrationServices.Server.ISServerExec.ISServerExecutionEvents.LoadPackage(PackageItem item)

   at Microsoft.SqlServer.IntegrationServices.Server.ISServerExec.ProjectOperator.StartPackage()

   at Microsoft.SqlServer.IntegrationServices.Server.ISServerExec.ProjectOperator.PerformOperation()"           
                          
"Package:Error: Package migration from version 8 to version 6 failed with error 0xC001700A ""The version number in the package is not valid. The version number cannot be greater than current version number.""."                   

"Package:Error: Error loading value ""<DTS:Property xmlns:DTS=""www.microsoft.com/SqlServer/Dts"" DTS:Name=""PackageFormatVersion"">8</DTS:Property>"" from node ""DTS:Property""."                                   

"Package:Error: The version number in the package is not valid. The version number cannot be greater than current version number."

The underlying cause is that currently (Sept 2014 timeframe) SSDT BI 2014 for Visual Studio 2013 does not support SSIS packages for SQL Server 2012.

It is a common feature request, and the product group is well aware of the demand.

https://connect.microsoft.com/SQLServer/feedbackdetail/view/944882/ssdt-bi-2014-backward-compatibility-for-ssis-2012

At present SSDT for VS 2013 only works with SQL Server 2014. You have to use SSDT-BI for Visual Studio 2012 with SQL Server 2012 SSISDB.

Thanks, Jason

On a clustered SQL Server 2008 R2 installation, we have installed SQL Server in two separate instances on each cluster node.

One Analysis Services resource will start, but the second one will not come online.

The error is shown in the application event log and in the MSOLAP\LOG\msmdsrv.log

Error(s): event id: 27 - failed to initialize SQL SQM timer
The service cannot be started: The following system error occurred: An attempt was made to access a socket in a way forbidden by its access permissions.

The cause

We found the first Analysis Services instance is listening on all IP addresses to port 2383 on the active node, so the second new instance cannot listen on its default port 2383 on the same node.

We could fail the second instance to the opposite node and it would come online fine. On the first node, we checked to see if the 2383 port is already in use.

1. Find the PID (process idenfitier) for msmdsrv.exe:

Using task manager, look at the list of Processes.
In the view menu, select the columns to view, and click PID (Process ID) to show that extra column.
Check the box in task manager that says "Show processes for all users"
Now sort the list of processes by name. Locate msmdsrv.exe process.
Found PID 5228

2. Checked which ports the first working AS instance is listening on using Netstat and cross referencing that pid

netstat -ano > output.txt
In the output.txt text look for the PID of the msmdsrv.exe to see which ports its listening on. Search (Control+F) for the PID number.
TCP 0.0.0.0:2383 PID 5228
TCP [::]:2383 PID 5228
So this indicates that the first msmdsrv.exe instance (pid 5228) is listening on ALL IP addresses.

Shortcut tip: you can pipe (|) output from netstat into findstr tool to narrow down the output quickly.

netstat -ano |findstr 2383

netstat -ano |findstr 5228

3. The problem is when the second instance start up, it cannot bind to its own IP address on port 2383 because the first instance has that port taken already.

We need to limit each instance to only its dedicated IP address from the cluster resources and not all IPs on the node.

The default port in described in https://technet.microsoft.com/en-us/library/cc646023.aspx

Note the default port is 2838 for default instances, and this is retained as a setting in the OLAP\CONFIG\mdmsdsrv.ini from the shared disk.
To confirm, find the port tag <Port>0</Port>  in the msmdsrv.ini. Zero indicates dynamic port assignment, which is 2383 for default instances, and dynamic on named instances.

The resolution

Check each Analysis Services resource in the cluster manager, to view the properties.

On the General Tab note the setting "Use Network Name for computer name" See also https://support.microsoft.com/kb/198893

This setting was unchecked on the first working AS instance. This means that the resource perceives the node name and all IPs,  instead of just the cluster resource IP and cluster virtual name. 

The second failing AS instance was a victim of this setting on the first instance. This setting was causing the first AS instance to bind to the 2383 port on all IPs and preventing the new AS instance from coming online on the same node.

So we check the setting "Use Network Name for computer name" to limit the instance to see only its IP dependency and its virtual name.
Then restart the AS instance where the change was made - take offline. Bring online in the cluster manager.
That solved the problem.

Now both instances can coexist on the same node and by failed over successfully.

We can run Netstat -ano to see that each msmdsrv.exe process uses port 2383 but on different IP addresses.

An error occurred while opening the model. Click Details for more information.

The operation cannot be executed since the database with the name of <name>, ID of <ID> already exists in the detached state in folder '\\?\D:\MSAS11.InstanceName\OLAP\Data\Datababase_c7f3679b-1d5f-44e8-b6d0-25d2de10b7fc.0.db'. Either attach the database or delete the folder and retry the operation.

To resolve this error, view the properties of Analysis Services, and ensure that your default Data Directory folder is browsable.

On the Analysis Server Properties page, General tab, check the box to Show Advanced (All) Properties.

Note your DataDir – this the folder where the tabular databases files live.

My default directory is C:\Program Files\Microsoft SQL Server\MSAS11.InstanceName\OLAP\Data

Now note your AllowedBrowsingFolders. This is a pipe | character delimited list of all the folders that the service is allowed to browse and show files in. Without this, Analysis Services will just not enumerate the directories.

C:\Program Files\Microsoft SQL Server\MSAS11.InstanceName\OLAP\Backup\|C:\Program Files\Microsoft SQL Server\MSAS11.InstanceName\OLAP\Log\|C:\Program Files\Microsoft SQL Server\MSAS11.InstanceName\OLAP\Data\

The problem we found was that the DataDir value was not in the AllowedBrowsingFolders list, so we added it to the string - |C:\Program Files\Microsoft SQL Server\MSAS11.TABULAR\OLAP\Data\

This could easily happen if the Default DataDir was changed, but the AllowedBrowsingFolders was not updated to reflect the change.

Internals

To figure this out I ran a profiler trace to watch the Command Begin and Command End events in particular.

I observed this pattern when it works:

1. Observe the temporary database creation

When I make a new Tabular project project in SSDT, the Analysis Services service creates a temporary tabular database, which consists of a folder and an XML pointer file to the folder.  This database is not visible in Management Studio, but is visible in the storage underneath the data folder itself.

Under the DataDir folder, ie.  C:\Program Files\Microsoft SQL Server\MSAS11.InstanceName\OLAP\Data, I see two indicators of the new tabular project being staged.

• /TabularProject6_jason_9f18f33a-4391-4ed7-9c39-95b18307d082.0.db/ (folder)
• /TabularProject6_jason_9f18f33a-4391-4ed7-9c39-95b18307d082.1.db.xml (file to point to the folder)

2. Observe the detach XMLA

When I close the Visual Studio project, it issues a detach command to remove the temporary database. The files in the folder should remain.

At this time of the detach, the Analysis Services deletes the .xml file pointer to the database, but the database .db folder and its contents remains in tact.

• /TabularProject6_jason_9f18f33a-4391-4ed7-9c39-95b18307d082.1.db.xml (gets deleted)
• /TabularProject6_jason_9f18f33a-4391-4ed7-9c39-95b18307d082.0.db/ (folder remains)

3. Observe the attach XMLA

When I open the same project again from the recently used list, it checks for existing folders, and attaches the same existing database using XMLA for attach.

A. MDX query check for the existing database by running an internal stored procedure

 SystemGetSubdirs 'C:\Program Files\Microsoft SQL Server\MSAS11.TABULAR\OLAP\Data'

B.  XMLA is issued to Attach the existing database from the listing of folder detected

C. Runs query to see the tabular data.

SELECT TOP 1000 [Model].[$DimEmployee].[EmployeeKey], [Model].[$DimEmployee].[ParentEmployeeKey], [Model].[$DimEmployee].[EmployeeNationalIDAlternateKey], [Model].[$DimEmployee].[ParentEmployeeNationalIDAlternateKey], [Model].[$DimEmployee].[SalesTerritoryKey], [Model].[$DimEmployee].[FirstName], [Model].[$DimEmployee].[LastName], [Model].[$DimEmployee].[MiddleName], [Model].[$DimEmployee].[NameStyle], [Model].[$DimEmployee].[Title], [Model].[$DimEmployee].[HireDate], [Model].[$DimEmployee].[BirthDate], [Model].[$DimEmployee].[LoginID], [Model].[$DimEmployee].[EmailAddress], [Model].[$DimEmployee].[Phone], [Model].[$DimEmployee].[MaritalStatus], [Model].[$DimEmployee].[EmergencyContactName], [Model].[$DimEmployee].[EmergencyContactPhone], [Model].[$DimEmployee].[SalariedFlag], [Model].[$DimEmployee].[Gender], [Model].[$DimEmployee].[PayFrequency], [Model].[$DimEmployee].[BaseRate], [Model].[$DimEmployee].[VacationHours], [Model].[$DimEmployee].[SickLeaveHours], [Model].[$DimEmployee].[CurrentFlag], [Model].[$DimEmployee].[SalesPersonFlag], [Model].[$DimEmployee].[DepartmentName], [Model].[$DimEmployee].[StartDate], [Model].[$DimEmployee].[EndDate], [Model].[$DimEmployee].[Status], [Model].[$DimEmployee].[EmployeePhoto] FROM [Model].[$DimEmployee]

4. Contrast to the observations on the failing broken instance

On the broken computer, when we get the error, I notice the same query is run, but instead of ATTACH XMLA statement, instead a XMLA ALTER/Create  statement is used to generate a new database, and we see the error message.

The real cause is that when the list of browsable folders cannot see the default data folder, then this query runs without any error, but returns no results

Made a new MDX connection to the tabular Analysis Service instance and ran it:

 SystemGetSubdirs 'C:\Program Files\Microsoft SQL Server\MSAS11.TABULAR\OLAP\Data'

<no result!>

Therefore on the broken instance, Analysis Services perceives there is no data folder matching the tabular project name. It tries to use the databasename_username_GUID to create a fresh database to stage the tabular model. The .db folder already exists and was detached earlier, so this causes a name collision and the error. The error is not lying, but the confusing thing is that we don’t see it create/detach/attach sequence because its hidden behind the scenes in the magic implementation.

The operation cannot be executed since the database with the name of <name>, ID of <ID> already exists in the detached state in folder '\\?\C:\MSAS11.InstanceName\OLAP\Data\Datababase_c7f3679b-1d5f-44e8-b6d0-25d2de10b7fc.0.db'. Either attach the database or delete the folder and retry the operation.

The database will get loaded on demand when the first connection is made to that database.  The Vertipaq engine will read all the database files from the disk and use the RAM as a cache to populate the data structures in memory.

For example, when using Management Studio, when you connect to the tabular instance shortly after starting the service, it will query metadata from ALL databases. The SSMS window may become nonresponsive until the server replies with all the metadata requested.

One simple way to forcibly warm the cache ahead of time is to run a simple Discover command to query each database metadata.

I have also filed a design change request to request the Analysis Services product group to please add a feature to autoload some or all databases. This could be triggered internally by the service whenever it starts to save user effort from creative workarounds. If you like the idea file it on https://connect.microsoft.com/sql

A simple workaround is to run a cache warming query:

Make a new MDX query window in Management Studio and runthis kind of query to illustrate how a simple cache warmer would work. This isn’t practical, because by the time Management Studio has shown you the list of databases, it has already queries all the metadata and the data files are loaded in RAM.

SELECT * FROM $System.DBSCHEMA_CATALOGS 

Make a new XMLA command window and run:

<Discover xmlns="urn:schemas-microsoft-com:xml-analysis">
       <RequestType>DBSCHEMA_CATALOGS</RequestType>
       <Restrictions />
       <Properties />
</Discover>

Automation to open the databases:

1. ASCMD.exe

To automate calling those we could use ASCMD.exe, which is an old sample back from 2005 days.  
https://msdn.microsoft.com/en-us/library/ms365187(v=SQL.90).aspx

You could make a batch file that you run shortly after starting the service.

2. Schedule it

Maybe a SQL Agent job is sufficient to run the code on a schedule every couple of minutes. However, it is harder to trigger it to start whenever the service restarts. I can use a SQL Server Analysis Services Query job step type.

The job interface requires a database name, which is not very useful for the discover query we need to run, so that’s annoying, but at least its automated.

Note: that the default SQL Agent Service Account may not have access to Analysis Services databases to run any query. If you use a domain account for your Agent service account, you may add the domain account into the AS security to run the query;  but if the Agent service is started with the system account or the NT Service\SQLAgent$ account, you would need to create a credential and proxy and specify the “Run as” on the job step to run as another user, and give that credential/proxy account rights in Analysis Services.

3. Windows Task Scheduler

Similarly you could use Windows Task Scheduler to run a batch file or start a powershell script.

A. Note in the Windows Event Viewer events when your AS Tabular instance starts.

There is an Event 0 (289) from the source MSOLAP$InstanceName when my service starts up- that may be a good trigger to launch the powershell script.

Log Name:      Application
Source:        MSOLAP$TABULAR
Date:          7/22/2013 9:14:44 PM
Event ID:      0
Task Category: (289)
Level:         Information
Keywords:      Classic
User:          N/A
Description:
Service started. Microsoft SQL Server Analysis Services 64 Bit Developer (x64) CTP 11.0.3000.0.

B. Make a powershell script in notepad  to prime the Analysis Services tabular databases.

Save the script with the .ps1 extension in a folder which is easily accessible.

Mine may not be perfect, but it is simply to load the sqlps, connect to my analysis services tabular instance, and run a query. Change the server name and instance name to match your computer.

import-module sqlps
cd SQLAS\MyServername\MyInstanceName
Invoke-ASCmd -Query "SELECT * FROM $System.DBSCHEMA_CATALOGS"

Test your script in the powershell command first before proceeding.

Powershell C:\Scripts\CacheWarmer.ps1

C. Now use Task Scheduler to watch for this event and trigger the cache warm up routine.

Start > Run > MMC.exe. Add the Task Scheduler snapin. Connect to the local server.

D. Click the menu option in the Actions pane to Create Task…

Pick the triggering event at Application, your MSOLAP$InstanceName, Event ID 0

Have it start a program - powershell and run your cachewarmer.ps1 script (as an argument – use double quotes around the path if it has spaces)

E. You can Finish the wizard, and view the properties.

If you need to change the account which runs the routine, do so in the properties. The account will need rights within your Analysis Services to run the query.

F. Test the Task, by stopping and starting (restart) your tabular instance service.

G. Now check your task scheduler status and see if it ran.

You can close the windows and check the history after you restart the services a few times to test it.

In Microsoft Customer Service and Support we saw this from time to time, at this was normally fine, but in the case where the metadata changes significantly, such as by changing a data type or changing the order or number of columns, the Script Component would start to misbehave since it may touch the wrong column or wrong data type and leave the data flow buffers in an unexpected  state.

Pseudo code example of the problem

Time 1: Original SSIS package design with a Dataflow task having 3 columns: ColA ColB ColC
Script Component in that Dataflow: Col(ColA).value=1; Col(ColB).value=2, Col(ColC).value=3

Then later someone changes the SSIS package design programmatically to introduce new columns.
Time 2: change inputs metadata to ColA, ColC, ColD

Then the Dataflow Script Component could incorrectly do things like use the ordinal position of columns and change the values incorrectly.
Col(ColA).value=1; Col(ColC).value=2, Col(ColD).value=3

This, above could cause data corruption too since you might be writing wrong types. If you put extra numbers of columns, or less numbers of columns, there are some unintended side effects.

So let’s prevent corruption and unexpected values by giving a true error in SSIS 2012

Therefore, the red flag went off in my mind after seeing this a couple of times, and I requested the product team to investigate a way to be sure that SSIS Dataflow buffers match what the Script Component expected when it was designed with the VSTA code editor. The solution is to have a verification mechanism that remembers the original metadata and if the column metadata changes since the task, and raise an error MetadataCheckumMisMatch when the metadata is known to be changed/mismatched.

Error: "Script Source" failed validation and returned validation status "VS_ISBROKEN".

Note the error is misspelled in the product – yes I know.

Therefore, in SQL 2012, if you programmatically edit the dataflow task design and change the input or output metadata which is hooked up to a Script Component (as a source, as a transformation, or as a destination) then you may see this error and you will wonder how to fix the problem.

To recompile the Script Component programmatically in SSIS 2008 – R2 we have these posts to help us

It was not well supported to do Script Component programmatically in SQL 2008, since you could not compile the script dynamically- the only supported way was is to open BIDS and click the Script Editor button.

There was a way to show the VSTA IDE and save the contents, which could be done programmatically. We blogged the workaround in SQL 2008 days

• https://dougbert.com/blog/post/Recompile-VSTA-scripts-programmatically-in-SSIS.aspx
• https://dougbert.com/blog/post/adding-a-vsta-script-task-programmatically.aspx

The gist is showing the IDE window, and closing it again automatically saved and recompiled the script code into binary code that gets saves into the SSIS package tags.

comp.ShowIDE();
comp.CloseIDE();

How to Recompile a Script Component programmatically in SSIS 2012

In SQL Server 2012 SSIS the script design was reworked to use a newer approach using the VSTA 3.0 editor, so there are a few differences now in how to open the Script and effectively recompile it with the current metadata values. This code sample shows how to take an existing package (.dtsx in this approach) and loop over all tasks to find the dataflow tasks, and then find all the script components within, to then open and save the code fresh, so that the script code gets the current dataflow pipeline metadata (to avoid the metadata is out of sync error). The highlighted lines are the meaty bits of this code. I don’t claim to be a developer, so please watch out for any bugs I may introduce with this code sample.

This sample is a C# for Visual Studio 2010 .Net 4.0

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using Microsoft.SqlServer.Dts.Design;
using Microsoft.SqlServer.Dts.Runtime;
using Microsoft.SqlServer.Dts.Pipeline.Wrapper;
using Microsoft.SqlServer.Dts.Pipeline;
using Microsoft.SqlServer.VSTAHosting;
using Microsoft.SqlServer.IntegrationServices.VSTA;

/* References
* // General SSIS references
* C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.Dts.Design\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SqlServer.Dts.Design.dll
* C:\Program Files (x86)\Microsoft SQL Server\110\SDK\Assemblies\Microsoft.SqlServer.DTSPipelineWrap.dll
* C:\Program Files (x86)\Microsoft SQL Server\110\SDK\Assemblies\Microsoft.SQLServer.DTSRuntimeWrap.dll
* C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.PipelineHost\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SQLServer.PipelineHost.dll
* C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.ScriptTask\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SqlServer.ScriptTask.dll
* C:\Program Files (x86)\Microsoft SQL Server\110\SDK\Assemblies\Microsoft.SQLServer.ManagedDTS.dll

* // Script related references
* C:\Program Files (x86)\Microsoft SQL Server\110\DTS\PipelineComponents\Microsoft.SqlServer.TxScript.dll
* or C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.TxScript\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SqlServer.TxScript.dll
* C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.VSTAScriptingLib\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SqlServer.VSTAScriptingLib.dll
* C:\Windows\Microsoft.NET\assembly\GAC_MSIL\Microsoft.SqlServer.IntegrationServices.VSTA\v4.0_11.0.0.0__89845dcd8080cc91\Microsoft.SqlServer.IntegrationServices.VSTA.dll
*/

namespace RunFromClientAppCS
{
    class Program
    {
        static void Main(string[] args)
        {
            string pkgLocation;
            Package pkg;
            Application app;
            DTSExecResult pkgResults;

            pkgLocation =
              @"c:\folder\Package.dtsx";
            app = new Application();
            pkg = app.LoadPackage(pkgLocation, null);

            try
            {
                TaskHost th = (TaskHost)pkg.Executables["Data Flow Task"];

                Executables pExecs = pkg.Executables;

                foreach (Executable pExec in pExecs)
                {

                    TaskHost taskHost = (TaskHost)pExec;
                    Console.WriteLine("Executable name = " + taskHost.Name);
                   
                    //Test if the task is a data flow
                    if (taskHost.InnerObject is Microsoft.SqlServer.Dts.Pipeline.Wrapper.MainPipe)
                    {

                        //Cast the Executable as a data flow
                        MainPipe pipe = (MainPipe)taskHost.InnerObject;

                        // Loop over each object in the dataflow
                        foreach (IDTSComponentMetaData100 comp in pipe.ComponentMetaDataCollection)
                        {

                            Console.WriteLine("Found Component Name = " + comp.Name);
                           
                            // Find the Script Components
                            if (comp.Name=="Script Component")
                            {
                                //Recompile the the Script Compoentn
                                CManagedComponentWrapper compWrap = comp.Instantiate();
                                ScriptComponentHost scriptComp = (compWrap as IDTSManagedComponent100).InnerObject as ScriptComponentHost;
                               
                                if (!scriptComp.LoadScriptFromComponent())
                                {
                                    throw new Exception("Failed to load script information from the component");
                                }
                               
                                if (scriptComp.CurrentScriptingEngine.VstaHelper == null)
                                {
                                    throw new Exception("Vsta 3.0 is not installed properly");
                                }
                               
                                if (!scriptComp.CurrentScriptingEngine.LoadProjectFromStorage())
                                {
                                    throw new Exception("Failed to load project files from storage object");
                                }

                                if (!scriptComp.SaveScriptProject())
                                {
                                    throw new Exception("Failed to save project");
                                }

                                Console.WriteLine("Recompiled Success: " + comp.Name);

                                scriptComp.CurrentScriptingEngine.DisposeVstaHelper();
                               
                            }

                            Console.WriteLine("  Component Name = " + comp.Name);
                        }
                    }
                }
            }
            catch (Exception e)
            {

                Console.WriteLine(e.Message.ToString());

            }

            // Run the Package to make sure it works.
            pkgResults = pkg.Execute();
           
            Console.WriteLine("Package Execution Result = " + pkgResults.ToString());
            Console.ReadKey();
        }

    }
}

Note: a security Gotcha for the %temp% folder

There is a gotcha which isn’t immediately apparent.

If you try to run this C# code from within a service when the Windows user profile is not loaded, it may fail. The code uses the temp folder for the users profile. Therefore, if you run this code from an SSIS parent package to recompile scripts in a child package, let’s say from a SQL Agent job or from stored procedures, then it may very well fail.  The call SaveScriptProject() doesn’t throw an error, it just doesn’t recompile the code.

You can add the proxy account, or the service account to have permissions to the TEMP folder which is failing.

There are generally two environment variables %TEMP% and %TMP%, and two sets – one for the user, and one for the system. If a users profile is not loaded, as is the case when an SSIS job runs as a proxy account user, then the system TEMP may be used.

When the service account for SQL Server launches something, the child process may perceive some system default TEMP folder. If a full user profile is loaded, it would presumably pick the user’s own TEMP folder under C:\Users\<AccountName>\Appdata\Local\Temp.    In theory, you could also change the TEMP pointer to point to a different folder if your system admin pushes back. That is the System variables in this case, but that is indeed global and affects all services on the box.

I’m am going to talk about network traffic encryption (incoming queries, results sent to clients like Management Studio and Excel, and writeback values sent from client to Analysis Services) and  not storage encryption (ie. database files and hard drive) here. I’m ignoring storage encryption in this blog post since it is another beast, and unless you have physical access to the hard drive or backup tapes, it doesn’t matter as much in todays secure data centers. .

Option 1: Tell Windows to Encrypt everything in the TCP traffic on your internal network

You could require Windows to  encrypt all network traffic on computers using windows features called IPSec and AuthIP. That’s outside my scope here, since that is complex to set up, and is usually a company wide effort as part of a security strategy, but its worth mentioning, since it means you encrypt traffic at a lower level and can secure systems more holistically.

External references:

For example, when running a job to do synchronization from a SQL Agent job, you may see the following error:

Executed as user: domain\SQLAgent.

Microsoft.AnalysisServices.Xmla.XmlaException: Either the ‘domain\SQLAgent’  user does not have permission to restore the 'MyDB' database, or the database already exists. Backup and restore errors: An error occurred while synchronizing the ‘MyDb’ database.

at Microsoft.AnalysisServices.Xmla.XmlaClient.CheckForSoapFault<XmlReader reader, XmlaResult xmlaResult, Boolean throwError) at Microsoft.AnalysisServices.Xmla.XmlaClient.CheckForErrormlReader reader. XmlaResult xmlaResult, Boolean throwliError)

at Microsoft.AnalysisServices.Xmla.XmlaClient. SendMessage(Boolean endReceivalliException.Boolean readSession, Boolean readNamespaceCompatibility)

at Microsoft.AnalysisServices.Xmla.XmlaClient.SendMessageAndReturnResult(String& result. Boolean skipResult)

at Microsoft.AnalysisServices.Xmla.XmlaClient.Execute(String command. String properties. String& result. Boolean skipResult. Boolean propertiesXmllsComplete)

at Microsoft.SqlServer.Management.Smo.Olap.SoapClient.ExecuteStatement(String stmt, StatementType stmtType. Boolean withResults, String properties. String parameters. Boolean restrictionListElement, String discoverType. String catalog)

at Microsolt.SqlServer.Management.Smo.Olap.SoapClient.SendCommand(String command, Boolean withResults,String properties)

at OlapEvent(SCH_STEPX pStep. SUBSYSTEMX pSubSystem. SUBSYSTEMPARAMSX pSubSystemParams. Boolean IQueryFlag). The step tailed.

To add the permissions:

1. Open the Management Studio window, and connect to your instance of Analysis Services.

2. Right click the server name and view the Properties

3. View the security page, and note if there are any users already present

4. Click the Add button and enter the name of the account which needs to run the synchronization, such as domain\account

By the way…. if you are going to run synchronization, you may want to set up the scheduled job  in SQL Agent and on the job step use a Proxy Account. Then you can limit the jobs which use the credentials of the proxy account, and add that account as a Server Admin in Analysis Services so that you don’t open up the SQL Agent account so widely to run work in Analysis Services.

Here are the instructions on using Proxy Accounts

1. Make an appropriate windows account (local account or domain account) in Users and Groups, or ask the Domain Admin to do so.

2. Make a credential (a saved account+password) in SQL Server Database Engine. https://msdn.microsoft.com/en-us/library/ms190703(v=sql.105).aspx

3. Make a proxy account to map the credential to a certain kind of jobs. https://msdn.microsoft.com/en-us/library/ms189064(v=SQL.105).aspx

4. Change the job step which launches the XMLA command or the SSIS package which runs your synchronization to use the proxy account.

Error:

Setup is missing prerequisites:

- 32-bit version of Microsoft Office 2010

Go to https://go.microsoft.com/fwlink/?LinkID=216742

The resolution is to use the Service Pack 1 version (or later) of the SQL Server 2012 Data Mining Add-ins. This download is for 2012 SP1:

https://www.microsoft.com/en-us/download/details.aspx?id=35578

There is a way to hack it a bit by telling the installer that Office is installed, and skipping the check by doing this kind of command line:

MSIEXEC /I C:\temp\SQL_AS_DMAddin.msi OFFICE2010PRESENT=1

“The field that you are moving cannot be placed in that PivotTable area”

We were trying to use an Excel workbook to drop in dimension attributes that are sourced from data living in the PowerPivot for Sharepoint Farm. We had our ODC file pointing to Analysis Services (MSOLAP provider) and the data source as the HTTP address of the XLSX living in the sharepoint farm. This means we are using Excel Services to feed in data to the client’s Excel Pivot Table from the server side PowerPivot workbook.

I used the sample PowerPivot workbooks from here to do some trial and error https://www.microsoft.com/en-us/download/details.aspx?displaylang=en&id=102

I confirm the problem is by design. Here’s why…

You cannot use cube dimensional attributes in the Values area box in an Excel PivotTable. That is the way regular PivotTables work. You may have been spoiled by PowerPivot Pivot Tables which allow you to drop in dimensional data attributes from a hierarchy into the Values area.

If you need this in a regular PivotTable to do the same for a PowerPivot workbook, there is a workaround. You can create the implicit measures in the server side PowerPivot workbook to get the ones that you need to share with the read only users ahead of time.

The PowerPivot workbook when stored in PowerPivot for Sharepoint, and accessed via Excel Services over HTTP is much like a Tabular Cube being accessed via MSOLAP provider.

To quote this KB article Differences between OLAP and non-OLAP PivotTables in Excel: https://support.microsoft.com/kb/234700

Layout and Design Differences

Dimensions vs. Measures

When you are working with a PivotTable report based on OLAP source data, dimensions can be used only as row, column, or page fields. Measures can be used only as data fields. When you drag a dimension to the data field drop area, or a measure to the row, column, or page field drop area, you receive the following error message:

The field that you are moving cannot be placed in that PivotTable area.

When a PivotTable report based on OLAP source data is active, the PivotTable toolbar displays an icon next to each row of fields. The icon indicates where Excel will allow you to place the field in your PivotTable report. If the icon is darker in the upper left, then the field is a dimension that you can drag to the row, column, or page field drop areas. If the icon is darker in the lower right, then the field is a measure that you can drag to the data field drop area.

The confusion between classic Excel Pivot Tables and PowerPIvot Pivot Tables

There is a conceptual difference between a classic Excel OLAP-based pivot table and a PowerPivot pivot table.

When using your source workbook as a regular PivotTable in this manner, you can only use those fields that are measures. In your case, the only measures that you have designed is “Sum of MeasuresDistinct Count of patient”. This is one of the advanced modeling activities that PowerPivot supports that Excel pivot tables do not.

• In a PowerPivot pivot table, you can pull down any numeric measure.
• In a classic Excel pivot table, you can only use measures in the cube

In PowerPIvot, when you pull down a numeric and drop it into the Values area , we create what we call an “implicit measure” behind your back. Classic Excel pivot tables cannot do that kind of automatic implicit measure.

Possible Solution:

What I suggest that you do is to take your numeric fields and drop them into the Values area in the original PowerPIvot workbook which lives in the PowerPivot gallery in Sharepoint. Doing so will create the implicit measure for you.

Then when you walk through the process to refer from Excel Services regular PivotTable to the PowerPivot “cube” in sharepoint, you will see a new measure group that you can drag down in the Excel pivot table.

It will show the implicit measures and their aggregate type (Count of, Sum of, etc).

Example of the PowerPivot workbook:

=====================================

Let’s prove it gives the error by design, and try the workaround:

A> I can repro the problem like this:

1. Upload your favorite PowerPivot xlsx into the PowerPivot gallery in Sharepoint.

https://site/PowerPivot/PowerPivot Healthcare Audit.xlsx

2. Now make an Excel workbook which refers to data on the HTTP address of the PowerPivot workbook.

Let’s point the Excel piivot table to data in the server side PowerPivot HTTP via Excel Services.

The steps are listed in the following article under the heading “Import data from PowerPivot Workbook”.
https://technet.microsoft.com/en-us/library/gg399165.aspx

As the MSDN article said, refer to Analysis Services in the Data tab

Point to the HTTP address of the PowerPivot workbook in Sharepoint.

The PowerPivot’s internal cube will appear

Rename if desired, and finish the ODC connection.

3. We can insert a PivotTable report into the workbook when prompted.

4. We can add measures from the PowerPivot’s “cube” into the Aggregated Values pane.

5. We  can add Dimensional data into Rows or Columns

6. We cannot add Dimensional attribute data into the Aggregated Values pane.

B> Contrast that with an OLAP based PivotTable. The same error applies to any PivotTable pointing to Adventureworks Multidimensional OLAP. Dimensional data cannot be places in Values pane.

1. Make a ODC connection to a multidimensional Analysis Server (MSOLAP provider where the Data Source is an AS server ,not the http PowerPivot .xlsx workbook)

2. Create a Excel PivotTable based on that connection.

3. Try to move a dimensional hierarchy attribute into the Values box.

C. Let’s compare to a PowerPivot Pivot table where Dimensional Data can be dropped into the Values box.

1. Click the PowerPivot ribbon > Pivot Table button

2. Note the PowerPivot Field List is different – and has 6 area boxes rather than the 4 above.

In a PowerPivot Pivot table you can use Dimensional data in the aggregated Values box. The same field which fails in a regular Pivot Table (as shown above) can be added just fine.

So dropping a dimensional measure created the “Count of _____” implicit measure on your behalf.

D. Now Let’s try to resolve the  problem by making implicit measures in the PowerPivot living in Sharepoint.

In the original PowerPivot workbook in the Gallery, the BI Developer can designate which fields should be declared as implicit measures.

1. Open the PowerPivot workbook from the gallery.

2. Open the Workbook in Excel (make sure the PowerPivot for Excel add-in is installed)

3. Make a new sheet and using the PowerPivot ribbon, add a “PivotTable” (the PowerPivot kind) and use the dimensional values that you need to use later in Excel Services in the PowerPivot field list for Values to create the implicit measures, such as COUNT OF ____ and SUM OF ____ shown here.

4. Save the workbook, so the new implicit measures are added back into the PowerPivot for Sharepoint server.

5. Now on the client in a different regular workbook, opened from Excel Services, or from Excel 2010 client itself… make a regular Pivot Table and choose the HTTP connection to the PowerPivot  XLSX in sharepoint (http uses MSOLAP via Excel Services)

6. In this workbook, you can see the implicit measures which were defined automatically in the PowerPivot “cube” itself living in Sharepoint. I can use those implicit measures in the regular PivotTable as aggregated values.

Voila!