I was recently helping a customer with their R installation.  After R was installed successfully, I went to the SQL Server Management Studio and ran the script needed to configure external services

EXEC sp_configure  'external scripts enabled', 1
RECONFIGURE WITH OVERRIDE
GO
RECONFIGURE;
GO

After this script ran successfully, I closed SSMS, and restarted the SQL Server service, noting when I did so that it told me that it was also going to be restarting SQL Server Launchpad, which is installed as part of R Server, was also going to be restarted.  Everything looked good until I tried to run my first R command.  I went for the super generic script

EXEC sp_execute_external_script @language=N'R'
,@script=N'OutputDataSet <-InputDataSet'
,@input_data_1=N'SELECT 1 as CheckToSeeIfRIsWorking'
WITH RESULT SETS
(
(
[CheckToSeeIfRIsWorking] INT NOT NULL
)
);

Much to my surprise after this I received an error

Msg 39019, Level 16, State 1, Line 1
An external script error occurred:
Unable to launch the runtime. ErrorCode 0x80070490: 1168(Element not found.).
Msg 11536, Level 16, State 1, Line 1
EXECUTE statement failed because its WITH RESULT SETS clause specified 1 result set(s), but the statement only sent 0 result set(s) at run time.

I looked in the log files and didn’t find any errors.  I checked the configuration manager to ensure that I had some user ids configured in the configuration manager.  Nothing seemed to make any difference.  Looking online, the only error that I saw which might possibly be close was a different error message about 8.3 naming and the working directory.

Resolving Error Code 0x80070490 Unable to launch the runtime for R in SQL Server

When I saw this error, I first checked the various SQL Log files. I didn’t find any errors.  I checked the configuration manager to ensure that I had some user ids configured in the configuration manager.  The default 20 users were there as expected. Nothing seemed to make any difference.  Looking online, the only error that I saw which might possibly be close was a different error message about 8.3 naming and the working directory.  I ran into a similar problem with R and Power BI.  R natively doesn’t understand anything but directories which are only 8 characters long.  If there are more letters than this in a directory, which is always the case with the default SQL Server directories, internally the directory is read as the first 8 characters and a tilde(~). This particular server also had an installation of SQL Server 2014 on it, and if you looked at the filename, it was impossible to tell which one was which as the directory read like this C:\PROGRA~1\MICROS~2\MSSQL1~1.MSS\MSSQL\EXTENS~1    I created a new directory with short names D:\Rwork and changed the working directory to the new subdirectory Rwork by updating the rlauncher.config file and restarting the SQL Server Launchpad Service.  I was really hopeful that this would work.  But when I ran the external script listed above, it still did not work.  What had I done wrong?  I didn’t delete the previous working directory, which was in the default location C:\Program Files\Microsoft SQL Server\MSSQL13.MSSQLSERVER\MSSQL\ExtensibilityData . When SQL Server creates this subdirectory, it also includes subdirectories for all of the users.  When I created the new directory, it didn’t have those subdirectories added.  I copied the subdirectories from the old working directory to the new one and tried again.  This time, success!

I hope you find this post helpful, as I know I could have used it when I was trying to figure this out. If it saved you time too, drop me a line and let me know.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

What is the Reason why installing R creates 20 User IDs?

If you have installed SQL, you may have noticed that it creates twenty user ids as part of the installation process. To many people the automatic creation of a number for SQL Server User IDs is alarming, and they want to know what these IDs are for and who will be using them. The answer for who could be using those IDs, is just about anyone running R and they are going to be using the resources of the server to do it. If you are a DBA and want to figure out how to stop this, keep reading as I promise to tell you, after I provide some context about SQL Server and R internals.

R Server and Launchpad

When R Server is installed as part of SQL Server, one way you can check to see if it is installed is to look to see if the Launchpad service is running. When R code is running it does not run within SQL Server OS. It is by definition an external process and the Launchpad exe serves as a conduit between SQL Server and the space where R is running. If you want to know more about R and SQL Server Internals, this article I wrote for SQL Mag will provide a lot more details. Microsoft designed the Launchpad service so that other languages might someday also run as R does on SQL Server. It also supports a feature of R Server which I wrote about, context switching. Context Switching provides the ability for users to utilize Server memory instead of the memory on their computers for running R, and access is granted through the use of one of the twenty ids created when R is installed.

Launchpad Settings – Where the External Users are Referenced

There are many reasons why a DBA might want to not allow clients to access server memory as that will tax the server. Turning it off is relatively simple. Go to the SQL Server Management Console and select SQL Server Launchpad for the instance of SQL Server running R Server.

In the picture of the screen, the instance of SQL Server I have running R Services is in SS2016. Right click on the server and select Properties, then click on the Advanced tab. When looking at the number of external users allowed by default, the number might look familiar. The reason there are twenty User IDs created for R Server is because Launchpad allocates by default external twenty users to connect from SQL Server to run R. If you don’t want to allow external users to run on a server, you will need to prevent the users from connecting by not enabling them to run R. To run R, users need to have db_rrerole permissions. If they do not have that, they cannot run R. On the production server, it is probably best that this permission not be granted to non-system users.

Since the External Users created are used by SQL Server when running R, it is not possible to set the number of external users to 0 as the Launchpad Service will not run, and no R Code can be executed anywhere. If the number of external users is modified, Configuration Manager provides a prompt window as a restart is required. If the number of External Users is set to 0, the Launchpad Service will not start. When the Launchpad Service tries to start, it will generate Error 1053: The Service Did Not Start in a Timely Fashion. The number of users has to be at least 1 for the service to be able to communicate with the external R components. If you add or reduce the number of External Users, the IDs will be either created or deleted to match the number listed.

Let me know if you found this information regarding SQL Server R Service information by commenting or messaging me on twitter. If you are interested in finding out more regarding the internals of SQL Server and R, you might be interested in reading this Article about the topic. I would also like to thank Bob Ward b | t of the Microsoft for helping me better understand the SQL Server R internals, and for patiently answering my questions on the topic.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

Whenever I talk about R, in the question and answer period I get a lot of questions from DBAs on how to monitor the impact of running R code on SQL Server. Fortunately Microsoft recently created a series of custom reports for SQL Server which provide a lot of useful information on how R is running on SQL Server. The custom reports are available on github here in a folder called SSMS-Custom-Reports. There are 8 different reports: R Services – Active Sessions, R Services – Configuration, R Services – Configure Instance, R Services – Execution Statistics, R Services – Extended Events, R Services – Packages, R Services – Resource Usage, R Services – Script. If it has been a while since you have added a custom report to SQL Server, this link should help you get the reports installed.

R Services Reports

The titles for the reports provide a good indication of what they do. Active Sessions will show how many users currently have their compute context set to run on the server. Configuration shows the current R Configuration, which is very useful information for checking on the R components on the server.

The report R Services – Configure Instance is for configuring R for the first time. If you have R setup already, chances are you won’t need this report, as it is really not a report but a script you need to run to configure R. R Services – Execution Statistics is a really neat way of monitoring the ongoing use of R. It keeps track of every time R is run on the server. If you run an R job then refresh the report, the execution count and time will go up.

R Services – Extended Events is also not a report but a list of all the extended events that are available for R Services. This is a handy bit of information, which can be a great reference tool for extended events monitoring. R Services – Packages lists the packages which are currently installed on SQL Server. When people write R, many lot of different packages are used within the script. Prior to running a package, check the information on this report to ensure the libraries used are installed on SQL Server. If the library is missing the code will not work. R Services – Resource Usage is a great way to see how R has been configured to run on the server. Notice I have created an External Pool for R. This is a configuration recommended by Microsoft to better monitor your R Services.

The last report R Services – TSQL Script is a bit of a mystery. The screen contains this line of text. This is the T-SQL script for generating the data in report: No Report . I don’t know what was the intent of the report. The Readme.md in github provides this documentation about the reports.
SSMS Custom Reports

Custom reports for SQL Server R Services built using SQL Server Report Builder. The reports can be added to SQL Server Management Studio to view various configuration and runtime information about R scripts execution in SQL Server.

That doesn’t provide any clues to me as to what the intent of the report is, so you may want to skip installing it altogether. If anyone does figure out how this report can be used, please drop me a line as I am curious. I was really excited when I found these reports and I think that they provide a lot of valuable insight for monitoring R on SQL Server 2016.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur

While R is an open source language, there are a number of different versions of R and each handles memory a little differently. Knowing which version is being used is important, especially when the code is going to be migrated to a server. As part of a SQL Server implementation, there are three different versions of R which come into play. The first is standard open source R, commonly known as CRAN R. This is the standard open source version of R which runs code in memory and is single threaded. The next version which will be installed as part of a SQL Server Installation is Microsoft R Open. This version of R was written to take advantage of the Intel Math Kernel Libraries [MLK]. Using the libraries speeds up many statistical calculations which use matrix operations. It also adds multi-threading capability to R as the rewrite provides the ability to use all available cores and processors and process in parallel. More information on how it works and how much faster Microsoft R Open is compared to standard R is available here. To use Microsoft R Open, once it is installed, in Rstudio should automatically start using it. To check out what version of R that is in use, within R Studio, go to Tools->Global Options and look at the R version.

R studio here is using Microsoft R Open. To take a look at the version in Visual Studio which has R Tools installed, go to Rtools and Options. Microsoft R Open is open source and can be used by anyone to speed up the execution of their R code. The version of R which is implemented within SQL Server is not this version of R. The R server version is proprietary, and Microsoft is not giving it way for free, as to run it on a server you need to purchase either SQL Server 2016 or R Standalone. That version is R Server. There is full compatibility on all of the versions of R. Code written in either CRAN R or Microsoft R Open will work in R Server. To write code for R Server, the R Client needs to be installed. The Visual Studio Screen shows this version of R installed on a PC which also is running SQL Server 2016 Developer edition.

R and Memory Consumption

One of R’s strengths and weaknesses is the fact that R runs in memory. This is good thing because it means R is very fast. It is not such a good thing when you need more memory than your machine has. When reading through the product information surrounding Revolution Analytics’ version of R [Revo R] which Microsoft bought, one of the features is the ability of this version of R to be able to use not only memory but disk, by applying chunking logic. The code is broken into smaller pieces, executed then put back together. Using Revo R, it’s possible to do this by using the specifically designed functions which all start with rx which implement this functionality, known as ScaleR. This does not mean that all R code running on SQL Server will be using these functions. Chances are most of it won’t, especially if the R code is being migrated from the existing data science team. Prior to being bought by Microsoft, Revolution Analytics charged money for their product. If your organization didn’t have a license, chances are no one was writing rx Functions. For people who don’t know anything about R, and have been asked to create packages to run R, there is an easy way to tell if the code has been written to use the chunking. If there are no words in the code that start with rx, like rxSummary, rxCube or rxHistogram to name a few, the code will not use chunking.

Yours Always

Ginger Grant

Data aficionado et SQL Raconteur