SQL Server 2016 Release Notes
Updated: December 3, 2015
Applies To: SQL Server 2016 Preview
Need help? MSDN forum | stackoverflow | Log an issue or suggestion at Microsoft Connect
This topic describes limitations and issues with SQL Server 2016
Try it out: » To download SQL Server 2016 Community Technology Preview 3 (CTP 3.1), go to Evaluation Center. » Have an Azure account? Then go Here to spin up a Virtual Machine with SQL Server 2016 already installed. |
For information on what's new, see What's New in SQL Server 2016, November Update.
Analysis Services upgrade
CTP 3.1 includes a partial fix for the upgrade bug in the previous CTP. Upgrading from 3.0 to 3.1 is no longer blocked by older Analysis Services client libraries, but upgrading from 2.x is still a problem. See Setup CTP 3.0 for recommendations on how to proceed.
Issue and customer impact: Consider the following limitations in CTP 3.1 when using large object (LOB) types ([n]varchar(max) and varbinary(max)) with memory-optimized tables and natively compiled modules:
Every value in a column of type [n]varchar(max) or varbinary(max) consumes at least 8 kilobytes of memory. This results in a large overall memory footprint for tables with small values in LOB columns.
The use of parameters and variables of type [n]varchar(max) or varbinary(max) with built-in functions LEN, LTRIM, RTRIM, and SUBSTRING and the concatenation operator '+' in natively compiled modules is not yet supported.
Workaround: No workaround.
Applies to: SQL Server 2016 CTP 3.1
After you upgrade from CTP 2.x to CTP 3.x, you have to reconnect databases that you previously configured for Stretch Database by running the stored procedure sys.sp_rda_reauthorize_db for each database. For more info, see sys.sp_rda_reauthorize_db (Transact-SQL).
Issue: You cannot enable Transparent Data Encryption (TDE) if you have one or more memory-optimized tables with LOB columns. Similarly, if you have a TDE enabled database, you cannot create a memory-optimized table with LOB column
Detection: You will get an error message
Issue: If TDE is enabled in a database that contains a columnstore index on memory-optimized table changing the database encryption key twice causes database corruption.
Detection: Database recovery will crash.
Workaround: None. We recommend you to disable TDE if you have a memory-optimized table with columnstore index
Applies to: SQL Server 2016 CTP 2.0, CTP3.0 and CTP3.1
Issue: This issue is resolved in CTP 3.1 and might still exist if you upgraded from CTP 3.0. In CTP 3.0, if you had a table with a clustered columnstore index (CCI) and a multi-column key nonclustered index (NCI), and you ran multiple updates to the same row, you might get inconsistent results in CTP 3.1 when you access the table through the NCI. You can get the correct results by accessing the table through the CCI.
Detection: To detect this index inconsistency run DBCC CHECKTABLE (<table-name>) with extended_logical_checks.
Workaround: Drop and re-create the NCI if it is corrupted.
Applies to: SQL Server 2016 CTP2.4 or CTP3.0.
Issue: This issue is resolved in CTP 3.1 and might still exist if you upgraded from CTP 3.0. In CTP 3.0, if your database had a columnstore index in the AlwaysON configuration with a readable secondary and you have offloaded the reporting workload using the columnstore index on the secondary replica, you could have data corruption on the secondary replica. This corruption could have occurred if you queried any of the columnstore index-based dynamic management views (DMVs) or catalog views on the secondary replica.
Detection: You can detect the corruption by running DBCC CHECKTABLE (<table-name>).
Workaround: If you detect the corruption and your database is not corrupt, first disable the readable secondary, and then restart the readable secondary. Perform a manual failover and then restart the new secondary replica. These steps are will remove a corrupt internal cached counter.
Applies to: SQL Server 2016 CTP3.0 and earlier.
This release adds several features and provides fixes for known issues in previous the previous CTP.
 Note |
|---|
Be sure to upgrade the SSAS tabular engine if you upgrade the CTP 3.1 tools. Schema changes to the tabular object model require that both engine and SSDT are from the same build. |
When creating a tabular model at compatibility level 1200, you can now create roles within the model itself in SSDT. This release removes a restriction previously called out in CTP 3.0.
You can now process or merge partitions for tabular databases at any compatibility level in SSMS.
You can now upgrade an 1100 or 1103 model to 1200 compatibility level, except when it contains pasted data. To upgrade, switch the compatibility level property to 1200 in either SSMS or SSDT.
Tabular models at the 1100 or 1103 compatibility level are fully functional with the exception of a few MDX limitations which are described below.
Several limitations called out in CTP 3.0 are now gone. In CTP 3.1, the list of known limitations is reduced to the following:
No session-scope MDX statements (named sets, calculated members, calculated cells, visual totals, default members, and so forth), but you can use query-scope constructs, such as the 'WITH' clause.
No drillthrough support.
The CTP 3.0 section below explicitly calls out several unsupported MDX features for Tabular 110x models, implying that they might be supported in the future. In fact, the following features that were not supported in the previous release will remain unavailable in SQL Server 2016:
Unqualified MDX member names are not supported in DirectQuery Tabular 110x models. When in DirectQuery mode, you need to specify the full object name in MDX to avoid name resolution errors. By full object name, an unqualified name like [Canada] would need to be expressed as [Customer].[Country].[Canada].
User-defined hierarchies are not supported in DirectQuery Tabular 110x models.
The following limitations still exist in CTP 3.1 for Tabular 1200 models but are expected to be lifted in a future CTP.
DirectQuery implementation for Tabular 1200 models. This will be a new set of behaviors that specifically target the 1200 compatibility level. This feature is not yet available in CTP 3.1.
AMO is still not supported for actions on objects below Database.
XMLA is still not supported for actions on objects below Database .
The Deployment wizard is still not supported.
PowerShell scripts are still not supported.
SSIS tasks and transforms against Tabular 1200 models are still not supported.
Setting reporting properties is still not supported .
Pasted tables are still not supported.
CTP 2.x-to-3.0 upgrade failure - Analysis Services
Analysis Services Management Objects (AMO) was refactored in CTP 3.0 into two assemblies, but the version number of the DLL was not updated accordingly. This causes Setup to load an existing, older version of AMO from the GAC rather than the new version provided in CTP 3.0, resulting in either one of these upgrade errors:
"An error occurred for a dependency of the feature causing the setup process for the feature to fail. Component name: SQL Server Analysis Services".
"Unable to connect to AS with name MSSQLServerOLAPService during upgrade"
These errors can occur even if you have a partial CTP 2.x installation of Analysis Services (for example, just the client libraries).
If you haven’t upgraded yet, you can avoid this error by doing either of the following options:
Option 1: Install AMO from the CTP3 feature pack before running upgrade from this link.
Option 2: Uninstall Analysis Services from the instance of SQL Server you want to upgrade. You can reinstall Analysis Services after the other components are upgraded.
If you ran upgrade and it failed, you can fix upgrade in the following ways.
When just Analysis Services fails
If upgrade fails for just Analysis Services, you can re-try upgrade and it will most likely succeed. Upgrade works on the second attempt because the new AMO version will have been registered on the first attempt, even though upgrade itself did not succeed.
For other failed components
Depending on which features are installed, you might see upgrade failures for multiple components. The workaround for this scenario is provided in both the Setup UI and in the Setup log file. The workaround varies depending on what failed. In most cases, the workaround is to roll back and then re-run upgrade. The command line for rolling back upgrade will include the uninstall action on an instanceID that is generated internally for upgrade purposes (this is MSSQLSERVER700 shown below). An example command line might look like this:
setup /q /action=uninstall /instanceid=MSSQLSERVER700 /features=RS,SQLENGINE,DQ,FULLTEXT,REPLICATION
| Note |
|---|
Installing CTP3 side-by-side with any other version of SQL Server has no known issues. |
Retirement of x86 Instance Features
The x86 (32 bit) version of SQL Server 2016 Setup no longer includes instance features. The following Instance Features are now available only in the x64 (64 bit) version of SQL Server Setup:
Database Engine
Analysis Services
Reporting Services - Native Mode
The x86 Setup is still available for installing tools and other shared Features.
Instance features installed from previous CTP releases of SQL Server 2016 will no longer be updated. It is recommended that previous CTP x86 installations are uninstalled and Instance Features are reinstalled using the x64 version of SQL Server 2016 Setup. Previous x86 installations with only Shared Features will continue to upgrade without issue.
Issue: Failover of an AlwaysOn Availability Group database replica which contains a table with a columnstore index, where queries were run on the secondary replica (readable secondary), may result in data inconsistencies on the table with the columnstore index. You can also run into this inconsistency if have queried any of the columnstore index based DMVs or catalog views on the secondary replica.
Detection: You can detect the data inconsistency by running DBCC CHECKTABLE (<table-name>) on the primary replica.
Workaround: If your scenario meets the criteria above and you have no consistency issues, you can avoid the issue by taking one of the these two actions:
Disable the readable secondary, restart the secondary replica, execute a manual failover, then restart the new secondary replica,
or
Upgrade to CTP 3.1.
This issue is resolved in CTP 3.1.
Applies only to: SQL Server 2016 CTP3.0 and earlier.
Issue: If TDE is enabled in a database that contains a columnstore index on memory-optimized table changing the database encryption key twice causes database corruption.
Detection: Database recovery will crash.
Workaround: None. We recommend you to disable TDE if you have a memory-optimized table with columnstore index
Applies to: SQL Server 2016 CTP 2.0, CTP3.0 and CTP3.1
Issue: During data synchronization you see the following error in the distribution agent log. The error message is incorrectly reporting error code 2812 when it should be error code 20598.
ErrorCode = '2812'
ErrorText = 'Could not find stored procedure 'sp_MSreplraiserror'.'
Workaround: Troubleshoot the issue as you normally would for error code 20598. For more information, see Error 20598 Details.
ErrorCode = '20598'
ErrorText = 'The row was not found at the Subscriber when applying the replicated command.'
Issue: On a table with a clustered columnstore index (CCI) and a multi-column key nonclustered index (NCI), multiple updates to the same row can give incorrect results when you access the table through the NCI. You can get the correct results by accessing the table through the CCI.
Detection:To detect this index inconsistency, run DBCC CHECKTABLE (<table-name>) with extended_logical_checks .
Workaround:You can take one of the two actions:
Enable TF-10210 globally by running DBCC TRACEON (10210, -1). Drop and re-create the index whenever it becomes corrupted.
This issue is resolved in SQL Server 2016 CTP 3.1. Upgrade to CTP 3.1 and then drop and re-create the NCI if it is corrupted.
Applies to: SQL Server 2016 CTP 2.4 or CTP 3.0.
Issue: You cannot create a columnstore index on an existing in-memory table. This restriction is only for CTP 3.0.
Workaround: To avoid this restriction, you can specify a columnstore index when you create the in-memory table.
Issue: The Analysis, Migrate, and Report (AMR) tool migrates disk-based tables to in-memory OLTP tables. This tool does not support migrating columnstore indexes.
Workaround:To migrate a columnstore index to an in-memory table, rename the existing table that has the columnstore index, create a new in-memory table with a columnstore index, copy the data to the new table, and then drop the old table.
Issue: If you try to enable Stretch Database on a mirrored database, the database engine will not raise a blocking message. However enabling Stretch on a mirrored database is not supported, and will be blocked in future releases.
Do not enable Stretch Database on a mirrored database.
Workaround: There is no workaround. Enabling Stretch on a mirrored database is not supported, and will be blocked in future releases.
If you discover a database credential with the name "StretchUser_DBM_<GUID>", you can safely drop the credential. This credential is no longer used. It was generated internally by Stretch Database in certain CTP releases of SQL Server 2016 prior to CTP 2.4.
Issue: The OPENJSON table-valued function is not supported in Hekaton native modules. However, OPENJSON function is not blocked in Hekaton stored procedures with the native_compilation attribute. If you try to create a native module with valid OPENJSON syntax, the statement will fail, a fatal error will be returned, the connection will be terminated, and the native module will not be created.
Workaround: If you want to use the OPENJSON function, you have to remove the NATIVE_COMPILATION attribute from the module and use interpreted stored procedures.
Issue: OPENJSON and the JSON built-in functions might use a lot of memory if they are executed on large results sets. Applying JSON functions on millions of records in the same query might consume a large amount of memory. As a result, the query might need to be manually canceled. In the worst case scenario, a query can take too much memory and make the SQL Server Database Engine unresponsive if there is not enough memory available.
Workaround: To work around this issue, use additional filters and indexes to reduce the number of rows in the result set to make sure that parsing JSON is not performed on a large number of rows with JSON text.
Issue: The OPENJSON function with an explicitly defined schema cannot return object and array values from JSON text. When a column in the WITH clause is mapped to a JSON property, only scalar properties can be returned. OPENJSON will return null if the JSON path in a column mapping references a JSON object or a JSON array.
The AS JSON option in a column definition that maps a column to a JSON array or JSON object is not supported.
Workaround: To work around this issue, rewrite queries to use OPENJSON without an explicitly defined schema, or use the JSON_QUERY function to retrieve JSON fragments.
The word DEFINITION was removed from column master keys. The CREATE COLUMN MASTER KEY DEFINITION statement becomes CREATE COLUMN MASTER KEY. sys.column_master_key_definitions is renamed to sys.column_master_keys. The column_master_key_definition_id column in sys.column_encryption_key_values is now column_master_key_id.
The COLUMN MASTER KEY argument of CREATE COLUMN ENCRYPTION KEY (Transact-SQL) has been renamed to COLUMN_MASTER_KEY.
SQL Server 2016 CTP2 Reporting Services is not supported in SharePoint integrated mode.
SQL Server 2016 Reporting Services (SSRS) SharePoint mode was only tested with SharePoint 2013.
DirectQuery is changing for tabular models for both existing and newly released 1200 compatibility levels, effectively resulting in two implementations of DirectQuery at each level. The following table describes availability in this release.
1200-level tabular models | Turned off for CTP 3.0. New behaviors will be announced for DirectQuery as they become available in upcoming CTP 3.x releases. |
1103-level tabular models | Operates under some of the same limitations as before -- namely, row-level security and calculated columns are still not supported. However, previous MDX limitations are slowly being lifted. Specifically, you can now issue MDX queries against an 1103 tabular model in DirectQuery mode, subject to the limitations described in the next note. |
The newest version of SQL Server Management Studio that aligns to CTP 3.0 cannot be used to process partitions for 1103 tabular models containing pasted tables. This will cause Management Studio to crash. Processing of Tables, Models and role management are still available, however.
For now, continue to use an existing version of Management Studio if you have tabular models with pasted tables.
In CTP 3.0, if you’d like to evaluate MDX data access for tabular 1103 models in DirectQuery mode, try to avoid the following known issues:
No session-scope MDX statements (named sets, calculated members, calculated cells, visual totals, default members, and so forth), but you can use query-scope constructs, such as the 'WITH' clause.
No unqualified MDX member names (use the full path). For instance, a query involving an unqualified name like [Canada] would produce a name resolution error, while a properly-qualified member name like [Customer].[Country].[Canada] is resolved as expected.
No native SQL queries (normally, Analysis Services supports a T-SQL subset, but not for DirectQuery models in this CTP).
No drillthrough.
No user-defined hierarchies.
No tuples with members from different levels in MDX subselect clauses.
Lastly, when generating MDX queries that contain custom calculations with many hierarchy and member navigation functions (such as PrevMember, Parent, ParallelPeriod), certain error conditions can generate mini dumps that do not provide descriptive error messages. If you see this problem, be aware that it’s a known issue that we expect to resolve soon.
To create a 1200-level tabular model, use the version of SQL Server Data Tools (SSDT) released in conjunction with CTP 3.0. You can view the compatibility level property of a model in the property pages of either Management Studio or in SQL Server Data Tools.
Building and using a 1200 level model is subject to the following limitations in CTP 3.0:
As noted, DirectQuery is disabled for 1200 models.
Upgrading the compatibility level from 1100 or 1103 to 1200 is not supported.
When processing partitions in Management Studio, the rowcount might inadvertently state that 0 rows are affected, even when processing is actually successful.
Roles in SSDT are disabled, but you can add roles in Management Studio.
Importing binary data is not supported.
AMO is not supported for actions on objects below Database for tabular models in 1200 compatibility level.
XMLA is not supported for actions on objects below Database for tabular models in 1200 compatibility level.
The Deployment wizard is not supported for tabular models in 1200 compatibility level.
PowerShell scripts are not supported for tabular models in 1200 compatibility level.
SSIS tasks and transforms against tabular models in 1200 compatibility level are not supported.
Setting Reporting properties is not supported.
Pasted tables are not supported.
SQL Server 2016 CTP2 Analysis Services SharePoint integration is not supported.
Issue: : When you click a slicer in a Power Pivot workbook that is hosted in a SharePoint document library, you see and error message similar to the following:
PivotTable Operation Failed: We Cannot locate a server to load the workbook Data Model”
And you see an error message similar to the following in the SharePoint ULS logs:
Uncaught CLR exception crossing the Interop boundary: Microsoft.AnalysisServices.Streaming.ServerNotFoundException: There are no servers available or actively being initialized.
A SQL Server 2016 Analysis Services server in Power Pivot mode cannot use the default instance name MSSQLSERVER. The Analysis Services installation will complete successfully, and you see no errors when you configure Excel Services “Data Model Settings” to use the Analysis Services server. However, Excel services is not able to use the Analysis Services server with the default instance name.
Workaround: Install a new instance of SQL Server 2016 Analysis Services in Power Pivot mode. On the Instance Configuration page of the SQL Server setup wizard, click Named Instance and type “POWERPIVOT”
More Information:
This is issue does not occur when upgrading an existing SharePoint 2013 and Power Pivot for SharePoint deployment to SQL Server 2016 Power Pivot for SharePoint . The issue does not occur because the Analysis Services instance is already a named instance “POWERPIVOT”.
Use the SQL Server Configuration Manager to verify if the Analysis Services instance name is the default name MSSQLSERVER.
If you use the new control flow template, you may encounter the following limitations and known issues.
Limitations
You can only create one executable in a control flow template. If you need multiple executables, use the Sequence container to wrap those executables.
You cannot run and debug the template in the designer.
Known issues
If you change the executable in the template from a container to a task, then when you reopen a package that contains an instance of this template, the designer may have trouble parsing the layout.
If you add a task and then a template instance to the control flow, the layout will break. To work around this issue, save the package before you add the first template instance.
Breakpoints, log providers, and event handlers that you add to a template instance are not persisted in the package and will be gone when you reopen the package in the designer.
If you install the SSDT Preview update for October 2015 side by side with a CTP release of SQL Server 2016 prior to CTP 3.0, then when you debug a package that contains a data flow task, the error "Unable to save to XML" is raised. To work around this issue, set the value of the Run64BitRuntime project property to false.
SQL Server 2016 CTP 2.4 is not supported in production environments.
Retirement of x86 Instance Features
The x86 (32 bit) version of SQL Server 2016 Setup no longer includes instance features. The following Instance Features are now available only in the x64 (64 bit) version of SQL Server Setup:
Database Engine
Analysis Services
Reporting Services - Native Mode
The x86 Setup is still available for installing tools and other shared Features.
Instance Features installed from previous CTP releases of SQL Server 2016 will no longer be updated. It is recommended that previous CTP x86 installations are uninstalled and Instance Features are reinstalled using the x64 version of SQL Server 2016 Setup. Previous x86 installations with only Shared Features will continue to upgrade without issue.
If you change the credentials for the remote Azure SQL Database used by an on premises database configured for Stretch, the on premises server may not be able to authenticate to the remote Azure SQL Database. As a result, all remote operations for Stretch Database will fail.
The SQL Server Data Tools installation is separate from the SQL Server installation. SQL Server Data tools can be installed standalone or it can be integrated into an existing Visual Studio installation. For more information, see Download Latest SQL Server Data Tools. The download does not currently include the Business Intelligence (BI) components.
BI: For BI scenarios, use the currently released version of SSDT-BI: Microsoft SQL Server Data Tools - Business Intelligence for Visual Studio 2013.
SQL Server 2016 CTP2 Reporting Services is not supported in SharePoint integrated mode.
SQL Server 2016 Reporting Services (SSRS) SharePoint mode was only tested with SharePoint 2013.
Issue: When you try to create Reporting Services subscriptions, the operation fails and you see an error message similar to the following:
library!ReportServer_0-6!1504!03/27/2015-13:02:35:: e ERROR: Microsoft.ReportingServices.Diagnostics.Utilities.ReportServerStorageException: An error occurred within the report server database. This may be due to a connection failure, timeout or low disk condition within the database. ---> System.Data.SqlClient.SqlException: Only members of sysadmin role are allowed to update or delete jobs owned by a different login.
This issue is caused when the Reporting Services Native mode windows service is configured to log on as NT Service account and also configured to connecting to catalog database with Service account as well.
Workaround: Complete one of the following workarounds:
Run Reporting Services under a Windows Account.
Change credentials that Reporting Services uses to connect to ReportServer database to either a Windows account or a SQL Server account.
More Information: The issues is similar to the KB article FIX: Error when you create new report subscriptions after you apply CU7 or CU8 for SQL Server 2012 SP1.
Applies to: Reporting Services Native mode only
SQL Server 2016 CTP2 Analysis Services SharePoint integration is not supported.
Issue: : When you click a slicer in a Power Pivot workbook that is hosted in a SharePoint document library, you see and error message similar to the following:
PivotTable Operation Failed: We Cannot locate a server to load the workbook Data Model”
And you see an error message similar to the following in the SharePoint ULS logs:
Uncaught CLR exception crossing the Interop boundary: Microsoft.AnalysisServices.Streaming.ServerNotFoundException: There are no servers available or actively being initialized.
A SQL Server 2016 Analysis Services server in Power Pivot mode cannot use the default instance name MSSQLSERVER. The Analysis Services installation will complete successfully, and you see no errors when you configure Excel Services “Data Model Settings” to use the Analysis Services server. However, Excel services is not able to use the Analysis Services server with the default instance name.
Workaround: Install a new instance of SQL Server 2016 Analysis Services in Power Pivot mode. On the Instance Configuration page of the SQL Server setup wizard, click Named Instance and type “POWERPIVOT”
More Information:
This is issue does not occur when upgrading an existing SharePoint 2013 and Power Pivot for SharePoint deployment to SQL Server 2016 Power Pivot for SharePoint . The issue does not occur because the Analysis Services instance is already a named instance “POWERPIVOT”.
Use the SQL Server Configuration Manager to verify if the Analysis Services instance name is the default name MSSQLSERVER.
SQL Server 2016 CTP 2.0 is not supported in production environments.
Side- by-side support is provided for the following SQL Server versions with SQL Server 2016 Community Technology Preview 3 (CTP 3.1).
SQL Server 2008
SQL Server 2008 R2
SQL Server 2012
SQL Server 2014
SQL Server 2016 CTP 1.4+
Upgrade to SQL Server 2016 Community Technology Preview 3 (CTP 3.1) is supported from the following SQL Server versions:
SQL Server 2008 SP4
SQL Server 2008 R2 SP3
SQL Server 2012 SP2
SQL Server 2014 RTM
SQL Server 2014 SP1
SQL Server 2016 CTP 1.4+
SQL Server 2016 Community Technology Preview 3 (CTP 3.1) is supported with the following operating systems:
Windows 8
Windows Server 2012
Windows 8.1
Windows Server 2012 R2
Windows 10
Windows Server 2016 Technical Preview
For more detailed information, see Hardware and Software Requirements for Installing SQL Server 2016
 Warning |
|---|
You can only upgrade to SQL Server 2016 Community Technology Preview 3 (CTP 3.1) Evaluation Edition from an earlier version of Evaluation Edition. If you are upgrading from a SKU to another SKU, you require a product key. |
Other
All Editions are supported: EVAL, DEV, ENT, ENTCORE, BI, STD, WEB, EXPR, EXPRWT, EXPRCOMP, EXPRSSMS
All SQL Languages are supported: ENU, DEU, JPN, CHT, CHS, KOR, FRA, ESN, ITA, RUS, PTB
All SKU's are Time Bombed; the Time Bomb is 180 days
WOW installs are supported.
SQL Server Books Online is available for the SQL Server 2016 Community Technology Preview 3 (CTP 3.1) release.
Issue: Managed reference content will not publish with CTP2 Books Online. Any links in conceptual topics in the SQL Server 2016 Community Technology Preview 3 (CTP 3.1) content that point to managed reference content will be broken links.
Workaround:
Refer to the content in the previous release, SQL Server 2014. Developer Reference for SQL Server 2014. For example:
More Information:
The following links point to some keys topics in the SQL Server 2016 Community Technology Preview 3 (CTP 3.1) Books Online content:
Issue: During data migration from a table configured for Stretch Database, locks on the table may be escalated from row level locks to table level locks. This may cause timeouts for INSERT or SELECT operations.
Workaround: If you detect issues caused by lock escalation, you can turn off lock escalation for the Stretch-enabled table by calling ALTER TABLE with the SET ( LOCK_ESCALATION = DISABLE ) option. Note that turning off lock escalation may result in high memory utilization that may cause behavior changes or issues elsewhere in the system. Evaluate and test this setting before making any changes.
More Information: For more info, see ALTER TABLE (Transact-SQL) and Lock Escalation (Database Engine).
Issue: On a database configured for Stretch Database, sp_spaceused cannot be used in an INSERT INTO EXEC statement if the value of the @mode parameter is N'ALL' or N'REMOTE_ONLY'. SQL Server doesn't support the nesting of INSERT INTO EXEC statements that results from this usage.
Workaround: None.
Issue: Dropping a database configured for Stretch Database may time out.
Workaround: Put the database in single user mode before dropping the database, as shown in the following example. For more info, see Set a Database to Single-user Mode.
ALTER DATABASE <database name> SET SINGLE_USER WITH ROLLBACK IMMEDIATE DROP DATABASE <database name>
More Information: In a database configured for Stretch Database, data is periodically moved from local tables to remote tables on a background thread. If you try to drop the database while this migration is happening in the background, the DROP operation may time out.
Issue: Stretch Database uses Distributed Query to access remote data in Azure SQL Database for a database that is configured for Stretch. Stretch Database connects to the remote data source by using SQL Native Client. SQL Native Client supports encrypted connections to SQL Server. Connections made by Stretch Database to access remote data are encrypted by using TLS. (For more info, see Connection Encryption and Certificate Validation.) However, the current implementation of Stretch Database does not validate the server certificate by default. Therefore the remote data access is susceptible to a man-in-the-middle attack. Forcing certificate validation as described in the workaround below may stop some applications that use Distributed Query to stop working. We anticipate that future releases of Stretch Database will require validation of the server certificate by default.
Workaround: You can mitigate the risk of a man-in-the-middle attack by forcing the encryption of client connections and the validation of server certificates. To check or change these properties, open the SQL Native Client 11.0 Configuration Properties in SQL Server Configuration Manager. Set the Force Protocol Encryption option to Yes and the Trust Server Certificate option to No on the computer on which you're using the Stretch Database feature.
More Information: Using Encryption Without Validation.
Issue: When Stretch Database is enabled, it monitors performance and quality by using telemetry data that it uploads asynchronously to Microsoft Azure. This telemetry is dependent on a stored procedure in Azure. Occasionally you may find entries in the local SQL Server error log with error number 2812 and the text "Could not find stored procedure" followed by the procedure name. This means that the telemetry infrastructure can't access the stored procedure due to a network issue or another unanticipated issue. If you can connect to Azure and the remote data for your Stretch-enable databases and tables, the cause of this error is typically related to the telemetry infrastructure. This error does not impact Stretch Database functionality and can be safely ignored.
Workaround: None.
Issue:Polybase, in CTP 2.0, supports only these server collations: SQL_Latin1_General_CP1_CI_AS (default US English collation) and Latin1_General_100_CI_AS_KS_WS.
Workaround: None.
Issue: To create a credential to be used with the PolyBase feature, the current implementation uses a version of the CREATE CREDENTIAL statement. However this is a temporary situation; in CTP3, the current functionality will be achieved using a new CREATE statement.
Workaround: None.
More Information: See Getting started with PolyBase and CREATE EXTERNAL DATA SOURCE (Transact-SQL)
Issue: For CTP2, data on Hadoop or Azure Blob storage is read-only.
Workaround: None.
Issue: Data corruption can occur when compressing a nonclustered-btree index on clustered columnstore index
Example: When compressing a delta rowgroup, if it needs to be broken into multiple compressed rowgroups due to either memory constraint or dictionary size limit, an internal structure used by nonclusterd btree indexes can get corrupt.
Workaround: A rebuild clustered columnstore index will fix the corruption with no data loss..
Issue: Data corruption issue with a nonclustered-btree index on clustered columnstore index.
Example: If you ALTER Column to change the datatype of a column, it can lead to data corruption after one or more delta rowgroups have been compressed.
Workaround: There is no work around. If you have a database backup available from before the corruption, you can attempt to restore the database. This issue was resolved in CTP 2.3.
Issue: When you create or rebuild a Nonclustered btree index, the process used is partially single threaded and therefore the process takes longer than expected. This is a regression from CTP1.4. This issue only affects the non-partitioned case. Index build plans for partitioned (aligned) non clustered indexes (NCI) over clustered column store indexes (CCI) are not affected by this issue.
Example : When you create or rebuild an NCI on a CCI.
Workaround: There is no work around.
Issue: Upgrading to CTP2 does not change read-only nonclustered columnstore indexes to be updateable, and pre-existing nonclustered columnstore indexes will have less functionality than they had before the upgrade. For example, after upgrading a nonclustered columnstore index, switching new partitions in and out will cause an error.
Workaround: After upgrading to CTP2, rebuild the nonclustered columnstore index.
More information: See ALTER INDEX (Transact-SQL) and use the REBUILD option.
SQL Server 2016 CTP2 Reporting Services is not supported in SharePoint integrated mode.
SQL Server 2016 Reporting Services (SSRS) SharePoint mode was only tested with SharePoint 2013.
Issue: When you try to create Reporting Services subscriptions, the operation fails and you see an error message similar to the following:
library!ReportServer_0-6!1504!03/27/2015-13:02:35:: e ERROR: Microsoft.ReportingServices.Diagnostics.Utilities.ReportServerStorageException: An error occurred within the report server database. This may be due to a connection failure, timeout or low disk condition within the database. ---> System.Data.SqlClient.SqlException: Only members of sysadmin role are allowed to update or delete jobs owned by a different login.
This issue is caused when the Reporting Services Native mode windows service is configured to log on as NT Service account and also configured to connecting to catalog database with Service account as well.
Workaround: Complete one of the following workarounds:
Run Reporting Services under a Windows Account.
Change credentials that Reporting Services uses to connect to ReportServer database to either a Windows account or a SQL Server account.
More Information: The issues is similar to the KB article FIX: Error when you create new report subscriptions after you apply CU7 or CU8 for SQL Server 2012 SP1.
Applies to: Reporting Services Native mode only
SQL Server 2016 adds a new Database Consistency Checker (DBCC) for tabular and multidimensional databases. However, there are known issues with this feature that won’t be fully resolved until CTP 3.0. For now, we recommend that you avoid using this feature.
SQL Server 2016 CTP2 Analysis Services SharePoint integration is not supported.
Issue: : When you click a slicer in a Power Pivot workbook that is hosted in a SharePoint document library, you see and error message similar to the following:
PivotTable Operation Failed: We Cannot locate a server to load the workbook Data Model”
And you see an error message similar to the following in the SharePoint ULS logs:
Uncaught CLR exception crossing the Interop boundary: Microsoft.AnalysisServices.Streaming.ServerNotFoundException: There are no servers available or actively being initialized.
A SQL Server 2016 Analysis Services server in Power Pivot mode cannot use the default instance name MSSQLSERVER. The Analysis Services installation will complete successfully, and you see no errors when you configure Excel Services “Data Model Settings” to use the Analysis Services server. However, Excel services is not able to use the Analysis Services server with the default instance name.
Workaround: Install a new instance of SQL Server 2016 Analysis Services in Power Pivot mode. On the Instance Configuration page of the SQL Server setup wizard, click Named Instance and type “POWERPIVOT”
More Information:
This is issue does not occur when upgrading an existing SharePoint 2013 and Power Pivot for SharePoint deployment to SQL Server 2016 Power Pivot for SharePoint . The issue does not occur because the Analysis Services instance is already a named instance “POWERPIVOT”.
Use the SQL Server Configuration Manager to verify if the Analysis Services instance name is the default name MSSQLSERVER.
Issue: In SQL Server Management Studio, under the Management\Extended Events node of an Analysis Services server, if you right-click and select New Session you see an error message. The details of the error message are different depending on what mode the Analysis Services server instance was installed.
Workaround: There is no work around to use the Analysis Services extended events in Management Studio. However you can use Extended events in the same way as past releases. For more information, see Use SQL Server Extended Events (XEvents) to Monitor Analysis Services and Tracing Analysis Services (SSAS) with Extended Events – Yes it works and this is how.
Issue: In SQL Server 2016 CTP2, the Incremental Package Deployment feature is not available in SQL Server Data Tools for BI (SSDT-BI).
Workaround: Use one of the following to incrementally deploy packages: Deployment Wizard, SQL Server Management Studio, stored procedures, and Management Object Model (MOM) API. See Deploy Packages to Integration Services Server for details.
Need help? MSDN forum | stackoverflow | Log an issue or suggestion at Microsoft Connect
