Clustering SSAS Tabular in real case studies

Hi,
I need to find more info about the Clustering of SSAS Tabular in real scenarios.
Any issues? Any particularities? Any warnings?
Tabular is on memory and not on storage.
Many thanks

The best practice is to put more nodes in load balancing - I think you can use Clustering, but when the service start on passive node in case of failover, you have to consider the time required to load data in memory from storage. I haven't seen a cluster
configuration for Tabular in production until now, only load balance configurations.
Marco Russo (Blog,
Twitter,
LinkedIn) - sqlbi.com:
Articles, Videos,
Tools, Consultancy,
Training
Format with DAX Formatter and design with
DAX Patterns. Learn
Power Pivot and SSAS Tabular.

Similar Messages

Need help ASAP - SSAS Tabular error has me stumped and others' solutions do not apply in this case.

Hello everyone,
Here's hoping no one will consider this to be the wrong forum. The data source is an Access database, but the issue is an SSAS Tabular issue. Kind of desperate here... the customer needs this ASAP, and I've been trying to resolve this for a week
already! Any help would be much appreciated!
A customer of mine (another employee of the ComIT department) is receiving an error trying to open an Access database via SQL Data Tools.
The error is "OLE DB or ODBC error: The Microsoft Access database engine cannot open or write to the file '\\[our_domain]\[the_path_to_the_file]\[database_name].accdb'. It is already opened exclusively by another user, or you need permission
to view and write its data.; 3051."
What perplexes me is that I do not receive this error, my coworkers on the DBA team do not receive this error, and our boss does not receive this error--even on the customer's machine. But the customer gets this error every single time.
The only difference is the credentials provided on the "Impersonation Information" window.
Here is our process:
We open SQL Data Tools, we create a new Analysis Services Tabular Project, we select the workspace server (our Analysis Services server... db5079\tabular) and test connection ("Test connection succeded") and click OK.
Then we click "Import From Data Source," select "Microsoft Access," enter the full network path to the database and test connection ("Test connection succeeded")
The next window says "Impersonation Information - Specify the credentials used by the Analysis Services server to connect to the data source when importing and processing data"
The credentials provided at this point is the only difference between me/my team/our boss and the customer. If we use our own credentials at this point--even on the customer's machine--we receive no error and everything is fine.
If we use the customer's credentials, we get the error above at the end of the next step (i.e., after we choose the data to import and then click Finish).
So that's it. On the same machine, the customer's credentials produce this error, and our credentials do not.
I have already added the customer as a server administrator to the Analysis Services server (db5079\tabular).
Copying the data source to another folder on the network or to his local machine produces the same results: his credentials produce the error, my credentials/my boss's credentials/etc. do not.
All of our machines are 64 bit, and the Analysis Services server is 2012 64-bit.
Please help!

UPDATE: As it turns out, all those who were able to import the data were local administrators on the Analysis Server (i.e., on the OS), and all those who were unable to import data were not.
When we added someone who couldn't import the data to the local Administrators group on the Analysis Server, they were able to import the data.
However, we can't give them local admin on the Analysis Server, and we are unable to determine what combination of user rights and permissions on folders we can grant the user as individual that will allow them to import the data.
Just as a test, we tried giving the user the same user rights that the admin group has, and the same permissions on all the drive as the admin group has, but that didn't work.
If it had worked, we could have started reducing the rights until we found the minimum necessary, but it didn't and once again we're stuck.
Please help!

Hide/show a column for some users in SSAS tabular

Hi,
I have a requirement to show one of the columns in a table only to some of the users. I know SSAS tabular only supports row level security. The only solution I came up with yet is to create two different models. Unfortunately this wastes a lot
of expensive RAM in the server. Is there a better way? Some kind of work-around? ;)
Best Regards
Ulf

One other way is to embed the security level into the table. You can do this by adding a "security_ID" into your structure and join it with a security table and join this to a security table which has a junction of Tablecolumns/User or TableColumns/UserGroups.
The next step is to use a Case When in your sql query and make a decision to show null or real data for any column.
Business Intelligence Architect

Excel SSAS Tabular error: An error occurred during an attempt to establish a connection to the external data source

Hello there,
I have an Excel report I created which works perfectly fine on my dev environment, but fails on my test environment when I try to do a data refresh.
The key difference between both dev and test environments is that in dev, everything is installed in one server:
SharePoint 2013
SQL 2012: Database Instance, SSAS Instance, SSRS for SharePoint, SSAS POWERPIVOT instance (Powerpivot for SharePoint).
In my test and production environments, the architecture is different:
SQL DB Servers in High Availability (irrelevant for this report since it is connecting to the tabular model, just FYI)
SQL SSAS Tabular server (contains a tabular model that processes data from the SQL DBs).
2x SharePoint Application Servers (we installed both SSRS and PowerPivot for SharePoint on these servers)
2x SharePoint FrontEnd Servers (contain the SSRS and PowerPivot add-ins).
Now in dev, test and production, I can run PowerPivot reports that have been created in SharePoint without any issues. Those reports can access the SSAS Tabular model without any issues, and perform data refresh and OLAP functions (slicing, dicing, etc).
The problem is with Excel reports (i.e. .xlsx files) uploaded to SharePoint. While I can open them, I am having a hard time performing a data refresh. The error I get is:
"An error occurred during an attempt to establish a connection to the external data source [...]"
I ran SQL Profiler on my SSAS Server where the Tabular instance is and I noticed that every time I try to perform a data refresh, I get the following entries:
Every time I try to perform a data refresh, two entries under the user name ANONYMOUS LOGON.
Since things work without any issues on my single-server dev environment, I tried running SQL Server Profiler there as well to see what I get.
As you can see from the above, in the dev environment the query runs without any issues and the user name logged is in fact my username from the dev environment domain. I also have a separated user for the test domain, and another for the production domain.
Now upon some preliminary investigation I believe this has something to do with the data connection settings in Excel and the usage (or no usage) of secure store. This is what I can vouch for so far:
Library containing reports is configured as trusted in SharePoint Central Admin.
Library containing data connections is configured as trusted in SharePoint Central Admin.
The Data Provider referenced in the Excel report (MSOLAP.5) is configured as trusted in SharePoint Central Admin.
In the Excel report, the Excel Services authentication settings is set as "use authenticated user's account". This wortks fine in the DEV environment.
Concerning SecureStore, PowerPivot Configurator has configured it the PowerPivotUnnattendedAccount application ID in all the environments. There is
NO configuration of an Application ID for Excel Services in any of the environments (Dev, test or production). Altough I reckon this is where the solution lies, I am not 100% sure as to why it fails in test and prod. But as I read what I am
writing, I reckon this is because of the authentication "hops" through servers. Am I right in my assumption?
Could someone please advise what am I doing wrong in this case? If it is the fact that I am missing an Secure Store entry for Excel Services, I am wondering if someone could advise me on how to set ip up? My confusion is around the "Target Application
Type" setting.
Thank you for your time.
Regards,
P.

Hi Rameshwar,
PowerPivot workbooks contain embedded data connections. To support workbook interaction through slicers and filters, Excel Services must be configured to allow external data access through embedded connection information. External data access is required
for retrieving PowerPivot data that is loaded on PowerPivot servers in the farm. Please refer to the steps below to solve this issue:
In Central Administration, in Application Management, click Manage service applications.
Click Excel Services Application.
Click Trusted File Location.
Click http:// or the location you want to configure.
In External Data, in Allow External Data, click Trusted data connection libraries and embedded.
Click OK.
For more information, please see:
Create a trusted location for PowerPivot sites in Central Administration:
http://msdn.microsoft.com/en-us/library/ee637428.aspx
Another reason is Excel Services returns this error when you query PowerPivot data in an Excel workbook that is published to SharePoint, and the SharePoint environment does not have a PowerPivot for SharePoint server, or the SQL Server Analysis
Services (PowerPivot) service is stopped. Please check this document:
http://technet.microsoft.com/en-us/library/ff487858(v=sql.110).aspx
Finally, here is a good article regarding how to troubleshoot PowerPivot data refresh for your reference. Please see:
Troubleshooting PowerPivot Data Refresh:
http://social.technet.microsoft.com/wiki/contents/articles/3870.troubleshooting-powerpivot-data-refresh.aspx
Hope this helps.
Elvis Long
TechNet Community Support

SSAS Tabular in DirectQuery - What are the workarounds for formula limitations?

Hello,
I need to create an SSAS Tabular model against the database of a live, real-time, line of business transactional system (i.e. a CRM).
The business requirement behind it is that we need to create some complex reports against live data, and our DW is only updated daily.
This live model will however be partitioned with a time-variance limitation (e.g. only records which are XX old can be returned).
Now here is the challenge. Since I am querying live data, then I believe the model must be configured in DirectQuery model. Am I right?
The issue is that DirectQuery mode is full of formula limitations. So my concern is, if I need a calculated column or measure that I cannot make it work due to DirectQuery limitations, then what are the alternatives?
Remember that the data source is from a live system, so it is not like I can create columns and measures in the underlying relational database.
Please advise.
Regards,
P.

Hi pmdci,
According to your description, you want to use some functions in calculated measure which are not supported in DirectQuery mode. Right?
In Analysis Services Tabular, since DirectQuery has the real time access and scalability, this comes with a price of restrictions on a number of DAX functions and missing Calculated Column feature. Generally the workaround for these scenarios
is replacing those functions with other functions which are supported in DirectQuery mode, or create columns in the data source. However, as you said, your environment is not possible to create columns in the database. And a lot of those limited
function are not replaceable, like time intelligence functions. So actually, there's no really effective workaround currently.
For you requirement, I suggest you submit Microsoft a feature request
at https://connect.microsoft.com/SQLServer
so that we can try to modify and expand the product features based on your needs.
Best Regards,
Simon Hou
TechNet Community Support

SSAS Tabular: Handling time datatypes

Folks,
I have two dimensions related to time in my DW. One is for dates (dDate), and another which is for the time of the day (dTime), with the granularity set to minutes (we don't care for seconds).
Here is the code for creating my dTime table:
CREATE TABLE dTime (
[timekey] smallint, --NOT SURE ABOUT THIS SEQUENCE AS THE PRIMARY KEY ** MUST REVIEW **
[Hour of Day (24)] tinyint, --0-23, military/European time
[Hour of Day (12)] tinyint, --1-12, repeating for AM/PM, for us American types
[Meridien] char(2), --AM/PM
[Minute of Hour] tinyint, --the minute of the hour, reset at the top of each hour. 0-59
[Half Hour] tinyint, --1 or 2, if it is the first or second half of the hour
[Half Hour of Day] tinyint, --1-24, incremented at the top of each half hour for the entire day
[Quarter Hour] tinyint, --1-4, for each quarter hour
[Quarter Hour of Day] tinyint, --1-48, incremented at the tope of each half hour for the entire day
[Time String (24)] char(5), --military/European textual representation
[Time String (12)] char(5), --12 hour clock representation sans AM/PM
[Time Value] time(0), --Time in 00:00:00 T-SQL time datatype
CONSTRAINT PK_dTime_timekey PRIMARY KEY CLUSTERED (timekey)
GO
For the sake of completion, I have recently added a time(0) field to my table. I am creating an SSAS tabular model based on data from this DW. However, it comes to my attention that SSAS in Tabular mode does not handle time-only fields. Is that correct?
Could someone please advise?
Best Regards,
P.

Multidimensional and Tabular are different.
Tabular has a dictionary for every column, so there is only one data type (datetime) that contain both. If you have a TIME data type in SQL, it is automatically imported in the right way and assigned to a date in the past (1/1/1900 or something similar).
In order to optimize the column, lower the number of unique values. If you don't need seconds, round to the minute.
Marco Russo (Blog,
Twitter,
LinkedIn) - sqlbi.com:
Articles, Videos,
Tools, Consultancy,
Training
Format with DAX Formatter and design with
DAX Patterns. Learn
Power Pivot and SSAS Tabular.

Case Studies for EJB Use

I am in the process of recommending J2EE/Weblogic as an architecture to a client,
but they have pushed back in wanting to know of successful deployments, and specifically
about the success of companies using EJB's instead of a simple Servlet approach.
Does anyone out there know of some good resources for case studies of companies deploying
J2EE/EJB's?

The question is: How complex and how transactional is the application? A
simple or a non-transactional application should probably be build without
EJBs.
In an enterprise where you have many apps built from shared components, EJBs
make sense too.
Peace,
Cameron Purdy
Tangosol, Inc.
Clustering Weblogic? You're either using Coherence, or you should be!
Download a Tangosol Coherence eval today at http://www.tangosol.com/
"Jason White" <[email protected]> wrote in message
news:3ce93a5f$[email protected]..
>
I am in the process of recommending J2EE/Weblogic as an architecture to aclient,
but they have pushed back in wanting to know of successful deployments,and specifically
about the success of companies using EJB's instead of a simple Servletapproach.
>
Does anyone out there know of some good resources for case studies ofcompanies deploying
J2EE/EJB's?

Capacity planning: case studies?

          Hello,
          does anyone know where i can find some case studies pertaining to clustering large-scale
          WLS/WLE applications. in particular, i am interested in the optimum number of
          servers
          in a cluster, nodes in a server instance, how to partition those nodes (by function/type?).
          thanks in advance,
          bob


Hi Bob,
          Unfortunately it depends primarily on the application. There's no simple
          math formula that provides what you are looking for. Generally, you can come
          to some conclusions only be load testing with reasonably good usage
          scenarios that reflect actual users. Then you can compare horizontal vs.
          vertical scale. Sometimes it is better to have a 32-processor box running
          one instance of WebLogic than to have a dozen 4-processor boxes because of
          little things like "db-is-shared".
          Peace,
          Cameron Purdy
          Tangosol Inc.
          Tangosol Coherence: Clustered Coherent Cache for J2EE
          Information at http://www.tangosol.com/
          "Bob Cleaves" <[email protected]> wrote in message
          news:3c45b7f7$[email protected]..
          >
          > Hello,
          >
          > does anyone know where i can find some case studies pertaining to
          clustering large-scale
          > WLS/WLE applications. in particular, i am interested in the optimum number
          of
          > servers
          > in a cluster, nodes in a server instance, how to partition those nodes (by
          function/type?).
          >
          > thanks in advance,
          > bob

OT - Case studies

I just wanted to get some insight and/or opinions on this. I
have seen a few designers out there that have what they call Case
Studies in their portfolios. Are these projects that fell through?
Are they fictitious projects?
I had talked to a friend of mine who told me that if I lose a
potential project or feel I need to add different styles to my
portfolio to just make up a company and design the site with fake
info...
What are other people's thought on this?

Well, I know my friend is a very ethical person with very
good morals and values. I do not think he was implying I should be
deceptive in my approach. He is a fairly successful application
developer and I don't think he would be as successful as he is by
being unethical in his business practices (although there was a guy
who had some little startup called Micro something that was
ethically questionable
I think his point was if I am having trouble building my
portfolio through pro-bono work with friends and family or a few
paid customer projects, then pick a category or type of business
and design a site as if I was hired to do so such as Real Estate.
Build it as if, so to speak but with readily apparent fake
information such 123 some street in somewhere with a fake email.
This way you are showing your skills and artistic style. I guess
almost like designing a template maybe.
Hence my original question of listing these type of sites in
the portfolio as case studies. That was what threw me off. I was
thinking I should put these type of sites in a case study group to
convey they were not real. When I think of a Case Study I think of
a lab assignment in school or like a controlled experiment. If that
makes sense.
Some other reasons for my initial question are that I
designed a site with a friend but he may be shutting it down and
not renewing his domain. Granted the site is nothing that fancy or
elaborate, but it was one of my first efforts and there are some
things in it that I was proud of. Once he removes this site from
his server, I would need to keep it on mine to show it to people
but some of it would not be functional anymore such as the email
and the url would no longer be his valid domain address. I also was
working on a design for a potential client that has pretty much
fallen through the cracks. I designed a nice concept logo and did a
quick 1 page photoshop mock up with my proposal. We never came to
an agreement on the price and the customer has just strung the
whole process along. I would like to finish this site and use it as
an example for the portfolio.
So I guess the real issue is if not calling these case
studies then what might be a better name for these examples that
conveys to people that they are not active sites but doesn't shout
out "Here is what I could've done but no one was willing to pay
me"

Slow Query Performance During Process Of SSAS Tabular

As part of My SSAS Tabular Process Script Task in a SSIS Package, I read all new rows from the database and insert them to Tabular database using Process Add. The process works fine but for the duration of the Process Add, user queries to my Tabular model
becomes very slow.
Is there a way to prevent the impact of Process Add on user queries? Users need near real time queries.
I am using SQL Server 2012 SP2.
Thanks

Hi AL.M,
According to your description, when you query the tabular during Process Add, the performance is slow. Right?
In Analysis Services, it's not supported to make a MDX/DAX query ignore the Process Add of the tabular. it will always query the updated tabular. In this scenario, if you really need good query performance, I suggest you create two tabular.
One is for end users to get data, the other one is used for update(full process). After the Process is done, let the users query the updated tabular.
If you have any question, please feel free to ask.
Regards,
Simon Hou
TechNet Community Support

SSAS Tabular model conection stops refreshing data for sometime

Hi,
I am using SSAS Tabular model and create connection files to use data with excel using powerpivot or powerview.
But recently for 1 or 2 months it stops working for few mins (generally 20-60) and starts working again.
It's happens for all 4 SSAS service on same server.
I am using AD roles with SSAS Tabular model to filter the data.
Everything is normal no event for this stop, no network error, no over usage of data.
ERROR is:
Errors in OLEDB provider. Could not connect to the redirector. Ensure that the SQLBrowser service is running on the 'ABC_Production' secver.

Hi pinaKPatel,
According to your description, you get the above error recently when using your tabular model. Right?
Based on the error message, firstly, please check if the OLEDB provider is installed properly. Then check the SQLBrowser service is running. If everything is working, it may caused by the port blockage. Please follow the steps below:
Use a fixed port and unblock that port in Windows Firewall. Connect to the server in this format: <servername>:<portnumber>.
Use dynamic port assignments and SQL Server Browser service. Unblock the port used by SQL Server Browser service in Windows Firewall. Connect to the server in this format: <servername>\<instancename>.
Use a fixed port and SQL Server Browser service together. This approach lets you connect using this format: <servername>\<instancename>, identical to the dynamic port assignment approach, except that in this case the server listens on a fixed port.
If the issue still persists, pelase redeploy your SSAS database.
Reference:
Errors in the OLE DB provider
If you have any question, please feel free to ask.
Simon Hou
TechNet Community Support

Evaluation of hw features of a machine hosting a SSAS Tabular model

Hi,
on a machine I've deployed a SSAS Tabular model that has two fact tables having each 200-300 millions of rows, and about 10-15 dimension tables. I've implemented some Excel workbooks, each of them queries one measure of one of the two fact tables.
Each workbook has 10-15 filters/slicers. I've tried some times to reduce the number of filter/slicer or of row expansion and I've registered a query response time from 5 to 15 seconds for a change selection for a filter/slicer. 15 seconds is a time too high
for the user requirements.
The formula for the two measures has some SUMs and DIVIDEs. In order to improve the query performance I've tried to write the formula in different manners and I've created partial calculations in the underlying dwh but with negligible performance improvements.
Cutting filters or slicers or row expansions is against the user requirements.
I've also captured some Profiler traces but with few indications.
I suspect that a possible issue derived from the hw features of the hosting machine. During the use of the pivot table the usage of memory is about 50-60% and the usage of CPU is 80-100%.
With the CPU-Z program I've captured the hw features:
I hope in any suggests deriving from the experiences and not from some papers, fe "Performance Tuning of Tabular Models in SQL Server 2012 Analysis Services", in order to improve the query performance from Excel connected to my tabular model.
Probably, the CPU features are not appropriate.
Many thanks

Hi Pscorca,
Based on your description, you want to if it is best practice for improve the query performance in hardware level by increase RAM or CPU. In this case, here are some document about it, please refer to the links below which might helpful for you.
Forcing NUMA Node affinity for Analysis Services Tabular databases
SSAS Tabular – NUMA and CPU Cores Performance
Regards,
Charlie Liao
TechNet Community Support

Data loaded to Power Pivot via Power Query is not yet supported in SSAS Tabular Cube

Hello, I'm trying to create a SSAS Tabular cube from a data loaded to Power Pivot via Power Query (SAP BOBJ connector) but looks like is not yet supported.
Any one tried this before? any workaround that make sense?
The final goal is pull data from SAP BW, BO Universe (using PowerQuery) and be able to create a SSAS Tabular cube.
Thanks in advance
Sebastian

Sebastian,
Depending on the size of the data from Analysis Services, one work around could be to import the data into into Excel and then make an Excel table and then use the Excel table as a data source.
Reeves
Denver, CO

SSRS Parameters using SSAS Tabular model get cleared

I have an SSRS report that uses data from a SSAS Tabular model. In the query designer, from the calendar dimension I choose a "Date Inclusive" filter and make it a parameter. I also choose to add another filter using and Organisation Unit
dimension and also make this a parameter. The report is written and deployed to a SharePoint 2013 library.
Most of the time, the report runs as expected with the parameters cascading off each other as expected. However, occasionally, parameters get cleared (either after changing a single value such as the Org Unit selection or sometime whilst the report
is being rendered). Sometimes you cannot select a value from the available values - you need to navigate somewhere else and then start over.
I changed the data source for the parameters to use SQL queries that return the same values as the MDX queries and the probably seems to have gone (time will tell)
This report has a child (detail) report that has one extract parameter. This parameter happens to have over 1,000 values. With the change of the parent report, you are now able to get to the child report. However, the child report seems
to exhibit the same problem with the parameters being cleared - and with a much higher frequency.
So, that leaves me wondering whether
anyone else has experienced this ?
is this an issue with SSRS 2012 and SSAS Tabular models (I have not seen this behaviour before and I have been using SSRS (since version 1) and SSAS Multi-dimensional (from when it was called "OLAP Services") ?

We applied SQL Server 2012 Service Pack 2 to the SharePoint farm (the SP Admin needed to re-create the service applications) and the problem is fixed

Filter Date Table (SSAS Tabular)

Hi Guys,
I'll try to define my issue as clear as possible.
Am creating a model for SSAS Tabular but I have following problem.
I have a Date Table that I want to filter depending on what the user select from another table with values like:
Today, Yesterday, This Week, Last Week, This Month, Last Month, .....
What I have tried:
(1) My Date table contains flags(columns) for each of these values. Filtering on these values is no problem. But then I have a long list of different possible flags.
(2) I created a copy of my date table but with the flags unpivoted (DateId, FlagName, Value (0 or 1))
I was hoping to be able to create somehow a relation between my date table and this table, but no success.
Also this way I would have a field that users can filter on (choose what period they want to see)
But sadly enough (2) did not work and I can not find any other way to find a solution for it.
Any help would be great.
Regards,
Sammy

In your date table, you can create different attribute for
each condition below:
1. Today, 2. Yesterday, 3. This Week, 4. Last Week, 5. This Month, 6. Last Month
So, if we consider above attributes there will be 6 different attributes with "Yes/No" Or "True/ False" Or "1/0" whatever flag.
I would suggest you to calculate these conditions in sql query which will be faster compared to calculated columns in the model.
Unfortunately, tabular model 2012 do not support dynamic set so you cannot build a single attribute with these values @server side.
I had similar question sometime back, check if you find something useful from this thread -
http://social.msdn.microsoft.com/Forums/sqlserver/en-US/302dd796-2677-44df-a76e-b053dcd14117/ssas-tabular-model-dynamic-fiscal-period?forum=sqlanalysisservices
If this post answers your query, please click "Mark As Answer" or "Vote as Helpful".

Clustering SSAS Tabular in real case studies

Similar Messages

Maybe you are looking for